AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Universum Prescription: Regularization Using Unlabeled Data

Xiang Zhang, Yann LeCun

AAAI 2017 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines · 28 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 dustbin-class universum prescription21-layer ConvNet · CIFAR-10, CIFAR-100 fine/coarse, STL-10 Tiny, ImageNet top-1/top-5test error (%): 7.02 -> 6.66; 37.58 -> 32.84; 22.74 -> 20.45; 31.16 -> 27.96; 34.39 -> 33.67; 13.68 -> 13.35Improved
add dustbin-class universum prescription21-layer ConvNet · STL-10test error (%): 31.16 -> 36.58Hurt
add uniform prescription21-layer ConvNet · CIFAR-100 fine/coarse and STL-10 Tinytest error (%): 37.58 -> 36.23; 31.16 -> 30.15Improved
add uniform prescription21-layer ConvNet · CIFAR-10, STL-10, ImageNet top-1/top-5test error (%): 7.02 -> 7.59; 31.16 -> 36.54; 34.39 -> 34.61; 13.68 -> 13.70Hurt
use a fixed-threshold background class instead of the baseline21-layer ConvNet · CIFAR-10, CIFAR-100 fine/coarse, STL-10, ImageNet top-1/top-5test error (%): 7.02 -> 8.38; 37.58 -> 40.57; 22.74 -> 24.97; 31.16 -> 38.95; 34.39 -> 34.69; 13.68 -> 13.84Hurt
increase the probability of sampling unlabeled data, up to a point (ablation)universum-prescription ConvNet · CIFAR-10, CIFAR-100, STL-10generalization gapImproved
increase the unlabeled-data sampling probability too far (ablation)universum-prescription ConvNet · CIFAR-10, CIFAR-100, STL-10test errorHurt
use unlabeled data that may overlap supervised classes rather than universum data (ablation)universum prescription on STL-10 · STL-10test error: 31.16 -> 36.58Hurt
use 80 million tiny images instead of STL-10's accompanying unlabeled data (ablation)universum prescription on STL-10 Tiny · STL-10 Tinytest error: 31.16 -> 27.96Improved
add dropout alongside universum prescription (ablation)universum-prescription ConvNet · CIFAR-10, CIFAR-100, STL-10test errorNo 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