AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Cognitive Deficit of Deep Learning in Numerosity

Xiaolin Wu, Xi Zhang, Xiao Shu

AAAI 2019 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded visual concept learning · 15 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
Test on triangles rather than training-set circlesSubitizing DCNN · Subitizing, trianglesaccuracy: 0.45Hurt
Test on squares rather than training-set circlesSubitizing DCNN · Subitizing, squaresaccuracy: 0.76Hurt
Swap foreground and background colorsSubitizing DCNN · Subitizing, swapped colorsaccuracy: 0.42Hurt
Use normalized boundary representations instead of raw image pixelsSubitizing DCNN · Subitizing across varied shapes, colors, and region-boundary representationsclassification accuracy: 1; 0.790; 0.920Improved
Scale objects up 50%Human-guided DCNN · Subitizing, scaled-up objectsprobability of judging 3 as 3: 0.496Hurt
Scale objects down 50%Human-guided DCNN · Subitizing, scaled-down objectsprobability of judging 3 as 3: 0.021Hurt
Add residual blocks to the subitizing network (ablation)Subitizing DCNN · Subitizing generalizationgeneralizationNo clear effect
Use morphology-based connectivity-preserving pruningDeterministic RCNN · Subitizing across object size and shapegeneralizationImproved
Train the deterministic RCNN from random weight initialization (ablation)Deterministic RCNN · Subitizingtraining convergenceHurt
Initialize weights slightly perturbed from the designed kernels (ablation)Deterministic RCNN · Subitizingtraining convergenceHurt
Decompose subitizing into erosion, connected components, and counting subnetworksSubitizing DCNN · Subitizing across object sizes and shapesgeneralizationImproved

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