AI Paper Atlas

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

CrossTransformers: spatially-aware few-shot transfer

Carl Doersch, Ankush Gupta, Andrew Zisserman

NeurIPS 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Fine-grained and cross-domain few-shot baselines · 57 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
replace global prototypes with spatially aligned local-feature comparisonsPrototypical Nets · Meta-Dataset Train-on-ILSVRCaverage rank: ProtoNets (Our implementation): 11.10; CTX: 3.85Improved
add 50% SimCLR episodes (ablation)CrossTransformers · Meta-Datasetaccuracy: OmniGlot: 76.52 -> 80.83; Aircraft: 79.65 -> 82.05; CU-Birds: 84.06 -> 82.01Mixed
add BOHB-style augmentation to CTX+SimCLR episodes (ablation)CrossTransformers · Meta-Datasetaccuracy: OmniGlot: 80.83 -> 82.21; Quick Draw: 68.84 -> 72.68; Aircraft: 82.05 -> 79.49Mixed
add test-time logistic regression on globally pooled features (ablation)CTX+SimCLR Eps+Aug · DTDaccuracy: 75.57 -> 80.29Improved
add test-time logistic regression on globally pooled features (ablation)CTX+SimCLR Eps+Aug · Aircraftaccuracy: 79.49 -> 77.41Hurt
add SimCLR instance-discrimination episodes rather than only SimCLR-style augmentation (ablation)Prototypical Nets · Meta-DatasetaccuracyImproved
apply SimCLR-style augmentation without instance discrimination (ablation)Prototypical Nets · Meta-DatasetaccuracyHurt
increase Prototypical Net capacity and image resolution (ablation)Prototypical Nets · Meta-DatasetaccuracyMixed
compute Batch Norm statistics from the test-time support set instead of using EMA (ablation)Prototypical Nets with SimCLR episodes · Meta-DatasetaccuracyImproved
increase feature-map resolution from 7×7 to 14×14 (ablation)CrossTransformers · Meta-Datasetaccuracy: Aircraft: 70.90 -> 79.65; CU-Birds: 80.29 -> 84.06; OmniGlot: 74.11 -> 76.52Improved
increase feature-map resolution from 7×7 to 14×14 (ablation)CrossTransformers · OmniGlotaccuracy: 87.26 -> 82.21Hurt

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