AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Implicit variance regularization in non-contrastive SSL

Manu Srinath Halvagal, Axel Laborieux, Friedemann Zenke

NeurIPS 2023 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Contrastive versus non-contrastive theory · 2 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
remove EMA target network (ablation)IsoLoss alpha=0.5 · CIFAR-10accuracy (%): 91.5 -> 91.5No clear effect
remove EMA target network (ablation)IsoLoss alpha=0.5 · CIFAR-100accuracy (%): 69.0 -> 64.3Hurt
remove EMA target network (ablation)IsoLoss alpha=0.5 · STL-10accuracy (%): 89.0 -> 87.4Hurt
remove EMA target network (ablation)IsoLoss alpha=0.5 · TinyImageNetaccuracy (%): 44.8 -> 40.4Hurt
remove EMA target network (ablation)DirectPred alpha=0.5 · CIFAR-10accuracy (%): 92.0 -> 12.1Hurt
remove EMA target network (ablation)DirectPred alpha=0.5 · CIFAR-100accuracy (%): 66.6 -> 1.6Hurt
remove EMA target network (ablation)DirectCopy alpha=1 · CIFAR-10accuracy (%): 91.3 -> 12.1Hurt
remove EMA target network (ablation)IsoLoss alpha=1 · CIFAR-100accuracy (%): 69.3 -> 63.0Hurt
replace DirectPred loss with IsoLoss (with EMA)closed-form predictor · TinyImageNetaccuracy (%): 40.1 -> 44.8Improved
replace DirectPred loss with IsoLoss (with EMA)closed-form predictor · CIFAR-100accuracy (%): 66.6 -> 69.0Improved
replace DirectPred loss with IsoLoss (with EMA)closed-form predictor · CIFAR-10accuracy (%): 92.0 -> 91.5No clear effect
use nonlinear predictor BYOL instead of IsoLoss (with EMA)IsoLoss · CIFAR-10accuracy (%): 91.5 -> 92.6Mixed

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