AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Learn to Preserve and Diversify: Parameter-Efficient Group with Orthogonal Regularization for Domain Generalization

Jiajun Hu, Jian Zhang, Lei Qi, Yinghuan Shi, Yang Gao

ECCV 2024 · Architectures and efficiency › Pruning and efficient tuning · approach: Adapter-based parameter-efficient tuning · 9 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 orthogonal regularization and a LoRA group (PEGO) on top of LoRALoRA on CLIP ViT-B/16 · DomainBed 5-dataset averageaccuracy: 75.0 -> 76.1Improved
add L_diversify to L_preserve (ablation)PEGO · OfficeHomeaccuracy: 82.95 -> 84.21Improved
add L_preserve to L_diversify (ablation)PEGO · PACSaccuracy: 96.37 -> 96.55Improved
add both orthogonal losses to an unregularized LoRA group (ablation)LoRA group · PACSaccuracy: 95.34 -> 96.55Improved
use a group of LoRA modules without regularization instead of single LoRA (ablation)LoRA · PACS and OfficeHomeaccuracy: 95.96 -> 95.34 (PACS); 83.41 -> 82.85 (OH)Hurt
use PEFT (LoRA, VPT, Adapter) instead of full fine-tuningCLIP ViT-B/16 · DomainBed 5-dataset averageaccuracy: 61.1 -> 75.0 (LoRA)Improved
increase LoRA rank above 8 (ablation)Baseline LoRA group vs PEGO · PACS and OfficeHomeaccuracyMixed
vary balancing coefficient from 1e-4 to 1e-1 (ablation)PEGO · PACS and OfficeHomeaccuracyImproved
PEGO versus zero-shot CLIP on TerraIncognitaCLIP ViT-B/16 · TerraIncognitaaccuracy: 31.3 -> 57.3Improved
PEGO versus zero-shot CLIP on PACSCLIP ViT-B/16 · PACSaccuracy: 96.8 -> 96.5Hurt
scale backbone to ViT-L/14 and apply PEGO instead of LoRALoRA on CLIP ViT-L/14 · TerraIncognitaaccuracy: 52.7 -> 57.2Improved
PEGO versus LoRA on PACS and VLCS with ViT-L/14CLIP ViT-L/14 · PACS, VLCSaccuracy: 98.1 vs 98.0 (PACS)No 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