AI Paper Atlas

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

Disentangling neural mechanisms for perceptual grouping

Junkyung Kim*, Drew Linsley*, Kalpit Thakkar, Thomas Serre

ICLR 2020 · Detection and segmentation › Open-set and few-shot detection · approach: Visual relationship and scene graph detection · 33 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
add horizontal feedback (ablation)recurrent CNN · Pathfinderaccuracy across difficulty levelsImproved
add top-down feedback (ablation)recurrent CNN · cABCaccuracy across difficulty levelsImproved
add both horizontal and top-down feedback (ablation)recurrent CNN · Pathfinder and cABCaccuracy across difficulty levelsImproved
remove horizontal feedback (ablation)TD+H-CNN · Pathfinderaccuracy as difficulty increasesHurt
remove top-down feedback (ablation)TD+H-CNN · cABCaccuracyHurt
remove both horizontal and top-down recurrence (ablation)TD+H-CNN · Pathfinder and cABCaccuracyHurt
use one processing timestep instead of recurrent processing (ablation)H-CNN and TD-CNN · Pathfinder for H-CNN; cABC for TD-CNNaccuracyHurt
use 8 timesteps instead of 1 timestep (ablation)fGRU recurrent models · Pathfinder and cABCperformanceImproved
replace fGRU modules with conventional GRUs (ablation)recurrent hierarchical architectures · Pathfinder and cABCperformanceHurt
replace fGRU modules with conventional LSTMs (ablation)recurrent hierarchical architectures · Pathfinder and cABCperformanceHurt
increase H+TD-CNN depth from 3 to 5 convolutional blocks (ablation)TD+H-CNN · Pathfinder and cABCperformanceNo clear effect
increase ResNet-18 width to match ResNet-50 parameter count (ablation)ResNet-18 · Pathfinder hardperformanceImproved

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