AI Paper Atlas

Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning

Graph Attention Prototypical Network for Robust Few-Shot Classification

Tingyun Liu, Licheng Liu, Qibin Zhang, Qiying Feng, C. L. Philip Chen

CVPR 2026 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-grained and cross-domain few-shot baselines

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 pseudo-label guided graph constructor and edge-aware graph attention (ablation)GAPNet · CIFAR-FS cleanaccuracy (%): 65.96 -> 75.71Improved
add pseudo-label guided graph constructor and edge-aware graph attention (ablation)GAPNet · CIFAR-FS 40% OOE noiseaccuracy (%): 53.34 -> 67.91Improved
add group attention broad learning module (ablation)GAPNet · clean datasetsaccuracy (%)Improved
add adaptive noise-robust prototype generator (ablation)GAPNet · noisy and clean settingsaccuracy (%)Improved
add SE block after ConvNet-4 (ablation)GAPNet · noisy and clean settingsaccuracy (%)No clear effect
increase prototype contrastive weight lambda from 0.01 to 0.1 or 1 (ablation)GAPNet · few-shot classificationaccuracy (%)Hurt
set fixed node rate above 0.5 in GABL (ablation)GAPNet · few-shot classificationaccuracy (%)Hurt
use 5 intra-class and inter-class nearest neighbours (ablation)GAPNet · few-shot classificationaccuracy (%)Improved
use GAPNet with same backbone and mean prototype as ProtoNetGAPNet · miniImageNet 40% in-episode noiseaccuracy (%): 40.57 -> 47.33Improved

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