AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Actively Testing Your Model While It Learns: Realizing Label-Efficient Learning in Practice

Dayou Yu, Weishi Shi, Qi Yu

NeurIPS 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Deep active learning · 0 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
ATL-NF importance-weighted test sampling instead of ARE quizzesATL-NF · MNISTestimation error x1e-3 (round 20): 8.31 -> 1.32Improved
ATL-NF instead of adapted AT integrateATL-NF · CIFAR10estimation error x1e-3 (round 20): 38.9 -> 6.20Improved
ATL-NF instead of adapted ASE integrateATL-NF · FashionMNISTestimation error x1e-3 (round 20): 8.42 -> 3.04Improved
add active feedback of test samples to training (ATL vs ATL-NF) (ablation)ATL · MNISThold-out test risk (round 20): 0.22 -> 0.19Improved
add active feedback (ATL vs ATL-NF) (ablation)ATL · FashionMNISThold-out test risk (round 20): 0.56 -> 0.51Improved
add active feedback (ATL vs ATL-NF) (ablation)ATL · CIFAR10hold-out test risk (round 20): 1.55 -> 1.53No clear effect
active feedback instead of random feedback (ablation)ATL · MNISThold-out test risk (round 20): 0.21 -> 0.19Improved
active feedback instead of random feedback (ablation)ATL · MNISTestimation error x1e-3 (round 20): 7.62 -> 1.87Improved
active feedback instead of random feedback (ablation)ATL · CIFAR10estimation error x1e-3 (round 20): 24.8 -> 6.57Improved
active feedback early in AL (round 4) (ablation)ATL · MNISTestimation error x1e-3 (round 4): 26.8 -> 14.6Improved
larger feedback size (83% vs 20%) (ablation)ATL · MNISThold-out test risk (round 20): 0.24 -> 0.20Improved
larger feedback size (83% vs 20%) (ablation)ATL · CIFAR10hold-out test risk (round 20): 1.63 -> 1.57Mixed

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