AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Deep Multitask Learning with Progressive Parameter Sharing

Haosen Shi, Shen Ren, Tianwei Zhang, Sinno Jialin Pan

ICCV 2023 · Architectures and efficiency › RNNs and normalization · approach: Representation geometry and collapse · 13 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
apply progressive parameter-sharing masks (MPPS) on top of hard-parameter sharingHPS · NYUv2Delta: +4.8% -> +6.5% (Gaussian), +5.7% (Bernoulli)Improved
apply MPPS on CityscapesHPS · CityscapesDelta: +0.7% -> +1.7%Improved
apply MPPS on Multi-CIFAR100HPS (VGG16) · Multi-CIFAR100average accuracy: 59.23% -> 67.62%Improved
apply MPPS to all layers instead of only last 6 (ablation)MPPS · NYUv2Delta: +5.7% -> -11.4%Hurt
restrict MPPS to last 6 layers rather than 12 or 18 (ablation)MPPS · NYUv2Delta: +4.5% (12 layers), +4.2% (18 layers) -> +5.7% (6 layers)Improved
add target distribution updating (ablation)MPPS · NYUv2Delta: +4.8% -> +5.7%Improved
progressive EC scheduling versus fixed highest exclusive capacity (Separ) (ablation)MPPS · NYUv2improvement over baselineImproved
add moderate frozen epochs (15, 25) to Linear scheduler (ablation)MPPS · NYUv2improvementMixed
combine MPPS with existing MTL methodsDWA, Auto-lambda, Graddrop, PCGrad, Cagrad, HPS, MTAN · NYUv2Delta: 6/9 methods improve, e.g. MTAN 10.3 -> 10.9; Random 7.2 -> 5.5; Uncert 4.7 -> 4.3Mixed
apply MPPS to smaller/larger backbones (ResNet-18, ResNet-34)HPS · NYUv2Delta: +2.7% -> +4.2% (ResNet-18); +5.7% -> +6.9% (ResNet-34)Improved
MPPS overhead on HPS trainingHPS · NYUv2, ResNet-50training time per epoch (s): 28.69 -> 39.34Hurt

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