AI Paper Atlas

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

DualCoOp: Fast Adaptation to Multi-Label Recognition with Limited Annotations

Ximeng Sun, Ping Hu, Kate Saenko

NeurIPS 2022 · Vision-language and multimodal › CLIP and zero-shot · approach: Open-vocabulary alignment from VLMs · 41 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 text supervision via CLIP-aligned dual prompts versus discrete label classifier (ablation)DualCoOp · MS-COCO partial labels 10%mAP: 70.6 -> 78.4Improved
text supervision in SST with few labels (ablation)SST · MS-COCO partial labels 10%mAP: 70.6 -> 69.1Hurt
learned prompts instead of hand-crafted pos/neg templates (ablation)DualCoOp · MS-COCO ZSLF1 @Top-3: 36.5 -> 50.3Improved
add negative prompt to positive learnable prompt (ablation)DualCoOp · MS-COCO ZSLF1 @Top-3: 44.5 -> 50.3Improved
halve prompt length from N=64 to N=32 (ablation)DualCoOp · MS-COCO ZSLF1 @Top-3: 50.3 -> 51.0No clear effect
class-specific region aggregation instead of multi-headed attention pooling (ablation)DualCoOp · MS-COCO partial labels 10%mAP: 72.1 -> 78.4Improved
increase test resolution 224 to 448 with attention pooling (train 224) (ablation)multi-headed attention variant · MS-COCO partial labels 10%mAP: 70.4 -> 65.9Hurt
increase test resolution 224 to 448 with class-specific aggregation (train 224) (ablation)DualCoOp · MS-COCO partial labels 10%mAP: 73.1 -> 76.0Improved
softmax-weighted aggregation vs average or max of regional logits (ablation)DualCoOp · MS-COCOmAPImproved

Later papers that compared with this method

6 later accepted papers report DualCoOp as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.

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