AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

A Sentence Speaks a Thousand Images: Domain Generalization through Distilling CLIP with Language Guidance

Zeyi Huang, Andy Zhou, Zijian Ling, Mu Cai, Haohan Wang, Yong Jae Lee

ICCV 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Recaptioning and text-supervised pretraining · 45 citations (OpenAlex)

Read the paper

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 CLIP distillation hint (KL on teacher logits) to ERM (ablation)ResNet50 · PACSaccuracy: 85.7 -> 88.4Improved
add absolute distance loss to ERM + Hint (ablation)ResNet50 · PACSaccuracy: 88.4 -> 89.0Improved
add relative distance loss on top of Hint + AD (ablation)ResNet50 · PACSaccuracy: 89.0 -> 89.4Improved
add AD + RD to ERM + HintResNet50 · DomainBed averageaverage accuracy: 72.5 -> 73.8Improved
add AD + RD to ERM + HintResNet18 · DomainBed averageaverage accuracy: 68.6 -> 69.6Improved
use CLIP text embedding instead of image embedding for AD supervision (ablation)ERM + Hint + AD · PACSaccuracy: 88.4 -> 89.0Improved
use 80-template prompt ensemble instead of a single prompt (ablation)ERM + Hint + AD · PACSaccuracy: 88.5 -> 89.0Improved
use cosine similarity for absolute distance instead of L1/L2/supervised contrastive (ablation)ERM + Hint + AD · PACSaccuracy: 88.0 / 88.1 / 88.6 -> 89.0Improved
use L2 vs KL for k1 in relative distance loss (ablation)ERM + Hint + RD · PACSaccuracy: 88.7 -> 88.8No clear effect
AD alone (text embedding only, no distillation hint) vs distillation hint (ablation)ERM · PACSaccuracy: 88.4 (Hint) vs 87.8 (AD)No clear effect
ensemble of students distilled from two different CLIP teachers (Mix Teacher) (ablation)RISE ResNet50 · PACSaccuracy: 89.4 -> 90.2Improved
ensemble two students from the same ViT B/16 teacher (ablation)RISE ResNet50 · PACSaccuracy: 89.4 -> 89.8Improved

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