AI Paper Atlas

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

Negative Margin Matters: Understanding Margin in Few-shot Classification

Bin Liu, Yue Cao, Yutong Lin, Qi Li, Zheng Zhang, Mingsheng Long, Han Hu

ECCV 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Embedding adaptation for few-shot classification · 397 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 negative margin to softmax loss (ablation)softmax pre-training · mini-ImageNet validation classes, 1-shotaccuracy: 45.98 -> 56.95Improved
add negative margin to cosine softmax loss (ablation)cosine softmax pre-training · mini-ImageNet validation classes, 1-shotaccuracy: 59.49 -> 63.68Improved
add negative margin to cosine softmax loss (ablation)cosine softmax pre-training · CUB validation classes, 1-shotaccuracy: 66.39 -> 69.17Improved
add negative margin to cosine softmax loss (ablation)cosine softmax pre-training · mini-ImageNet->CUB validation, 5-shotaccuracy: 61.99 -> 64.04Improved
add negative margin to softmax loss (ablation)softmax pre-training · CUB validation classes, 5-shotaccuracy: 80.21 -> 80.60No clear effect
add negative margin (ablation)Neg-Cosine ResNet-18 · mini-ImageNet novel, 1-shotaccuracy: 54.51 -> 60.25Improved
add weight decay on top of negative margin (ablation)Neg-Cosine ResNet-18 · mini-ImageNet novel, 1-shotaccuracy: 60.25 -> 62.21Improved
add DropBlock on top of negative margin and weight decay (ablation)Neg-Cosine ResNet-18 · mini-ImageNet novel, 1-shotaccuracy: 62.21 -> 62.33No clear effect
add label smoothing (ablation)baseline softmax · mini-ImageNet validation classes, 1-shotaccuracy: 59.49 -> 58.02Hurt
add label smoothing (ablation)baseline softmax · mini-ImageNet base classes, 1-shotaccuracy: 85.45 -> 88.93Improved
add negative margin (ablation)baseline softmax · mini-ImageNet validation classes, 5-shotaccuracy: 78.90 -> 81.60Improved
add negative margin (ablation)baseline softmax · mini-ImageNet base classes, 1-shotaccuracy: 85.45 -> 79.92Hurt

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