AI Paper Atlas

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

Model-Agnostic Boundary-Adversarial Sampling for Test-Time Generalization in Few-Shot learning

Jaekyeom Kim, Hyoungseok Kim, Gunhee Kim

ECCV 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Meta-learning analysis and benchmarks

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 MABAS test-time fine-tuningSTL · miniImageNet 5-way 1-shotaccuracy (%): 55.98 -> 60.19Improved
add MABAS test-time fine-tuningFSwF · miniImageNet 5-way 5-shotaccuracy (%): 69.94 -> 78.28Improved
add MABAS test-time fine-tuningMetaOptNet · miniImageNet 5-way 5-shotaccuracy (%): 78.55 -> 81.01Improved
add MABAS test-time fine-tuningMetaOptNet-Proto · miniImageNet 5-way 5-shotaccuracy (%): 78.36 -> 82.70Improved
add MABAS test-time fine-tuningSTL · CIFAR-FS 5-way 5-shotaccuracy (%): 83.62 -> 84.29Improved
add MABAS test-time fine-tuningFSwF · FC100 5-way 5-shotaccuracy (%): 49.75 -> 54.95Improved
add MABAS test-time fine-tuningMetaOptNet-Proto · FC100 5-way 1-shotaccuracy (%): 40.60 -> 42.31Improved
add boundary-adversarial samples instead of naive fine-tuning on support only (ablation)FSwF · miniImageNet 5-way 1-shotaccuracy (%): 55.73 -> 60.45Improved
add boundary-adversarial samples instead of naive fine-tuning on support only (ablation)FSwF · miniImageNet 5-way 5-shotaccuracy (%): 70.14 -> 78.28Improved
naive fine-tuning on support samples only (ablation)FSwF · miniImageNet 5-way 5-shotaccuracy (%): 69.94 -> 70.14No clear effect
add MABAS to ResNet-12 embeddingFSwF · miniImageNet 5-way 5-shotaccuracy (%): 69.94 -> 78.28Improved
add MABAS to Conv64 embeddingFSwF · miniImageNet 5-way 5-shotaccuracy (%): 71.89 -> 72.15No clear effect

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