AI Paper Atlas

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

An Ensemble of Epoch-wise Empirical Bayes for Few-shot Learning

Yaoyao Liu, Bernt Schiele, Qianru Sun

ECCV 2020 · Label-efficient and robust learning › Zero-/few-shot and meta-learning · approach: Normalization and unsupervised meta-learning · 149 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 E3BM epoch-wise ensemble with hyperprior learners to SIBSIB · miniImageNet 1-shotaccuracy (%): 70.0 -> 71.4Improved
add E3BM to SIBSIB · tieredImageNet 1-shotaccuracy (%): 72.9 -> 75.6Improved
add E3BM to SIBSIB · FC100 5-shotaccuracy (%): 55.9 -> 57.1Improved
add E3BM to MAMLMAML · miniImageNet 1-shotaccuracy (%): 48.70 -> 53.2Improved
add E3BM to MTLMTL · miniImageNet 1-shotaccuracy (%): 63.4 -> 64.3Improved
add E3BM to MTLMTL · FC100 5-shotaccuracy (%): 60.1 -> 60.5Improved
use transductive instead of inductive learning of hyperprior learners (ablation)MTL+E3BM (LSTM) · miniImageNet 5-shotaccuracy (%): 81.0 -> 81.1Improved
use transductive instead of inductive learning (ablation)MTL+E3BM (LSTM) · tieredImageNet 5-shotaccuracy (%): 85.0 -> 85.3Improved
use LSTM instead of FC hyperprior learner (ablation)MTL+E3BM (transductive) · miniImageNet 5-shotaccuracy (%): 80.7 -> 81.1Improved
use LSTM instead of FC hyperprior learner (ablation)SIB+E3BM · tieredImageNet 5-shotaccuracy (%): 83.8 -> 84.3Improved
use hyperprior-generated ensemble instead of 5 independent MAML learners (ablation)MAML · miniImageNet 1-shotaccuracy (%): 52.1 -> 53.2Improved
use FC hyperprior vs 5x MAML (same ensemble size) (ablation)MAML · miniImageNet 1-shotaccuracy (%): 52.1 -> 52.1No 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