AI Paper Atlas

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

Semi-Supervised Gaze Estimation via Disentangled Subspace Contrastive Learning

Qida Tan, Hongyu Yang, Wenchao Du

ICML 2026 · Label-efficient and robust learning › Semi-supervised learning · approach: GAN-based semi-supervised 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
add DSCL to supervised baselineResNet-18 baseline · Gaze360, 5% labelsMAE (degrees): 23.00 -> 14.21Improved
add DSCL to existing gaze methodsGazetr, RT-Gene, FullFace, Dilated-Net, ST-WSGE · Gaze360 (20/10/5% labels)MAE (degrees): avg 16.57 -> 15.14 (Gazetr), 17.08 -> 14.79 (ST-WSGE)Improved
add DSCL to cross-domain baselinesBaseline DG, PureGaze, PnP-GA · Gaze360 -> MPIIGaze / EyeDiapMAE (degrees): 18.75 -> 10.21 (baseline, EyeDiap 5%)Improved
remove Jacobian regularization LJ (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 17.66 -> 14.21Improved
add unsupervised ranking loss LUR (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 18.88 -> 14.21Improved
add unsupervised contrastive loss LUC (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 15.97 -> 14.21Improved
add supervised contrastive loss LSC (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 15.40 -> 14.21Improved
add both unlabeled-data losses (LUR and LUC) (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 21.72 -> 14.21Improved
add supervised initialization phase (Lreg + LJ) (ablation)DSCL · Gaze360, 10% labelsMAE (degrees): 19.96 -> 13.94Improved
use DSCL instead of other semi-supervised methodsResNet-18 baseline · Gaze360, 5% labelsMAE (degrees): CLSS 20.77 -> DSCL 14.21; UCVME 21.49; RankUp 24.95Improved
add external unlabeled data (WebFace and CelebA) (ablation)DSCL · Gaze360 -> MPIIGaze, 10% labelsMAE (degrees): 10.57 -> 9.15Improved
add WebFace only as external unlabeled data (ablation)DSCL · Gaze360, 5% labelsMAE (degrees): 14.21 -> 14.26No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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