AI Paper Atlas

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

Position-Guided Text Prompt for Vision-Language Pre-Training

Jinpeng Wang, Pan Zhou, Mike Zheng Shou, Shuicheng Yan

CVPR 2023 · Vision-language and multimodal › CLIP and zero-shot · approach: Adapting VLMs for zero-shot classification · 33 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 position-guided text prompts to ViLTViLT · MSCOCO 5K test, zero-shot image-to-text retrievalR@1: 41.3 -> 55.1Improved
Add position-guided text prompts to BLIPBLIP · Flickr30K 1K test, zero-shot retrieval average recallAvg: 81.7 -> 88.4Improved
Add position-guided text prompts to BLIPBLIP · NoCaps validation overallCIDEr: 98.8 -> 106.0Improved
Add position-guided text prompts to ViLTViLT · MSRVTT 1k test, text-to-video retrievalR@1Improved
Use an additional position-prediction pretext task rather than only the prompt (ablation)BLIP · COCO captioningCIDEr: 121.2 -> 123.5Improved
Use the text prompt instead of the additional pretext task (ablation)BLIP · COCO captioningCIDEr: 123.5 -> 127.2Improved
Mix different prompt templates rather than use the standard block prompt (ablation)BLIP · COCO retrievalTR@1: 73.2 -> 72.3Hurt
Add object tags, prompt, and position together rather than using object tags alone (ablation)BLIP · COCO retrievalTR@1: 70.2 -> 73.3Improved
Add object tags, prompt, and position together rather than omitting position (ablation)BLIP · COCO retrievalTR@1: 70.3 -> 73.3Improved
Add object tags, prompt, and position together rather than omitting the prompt (ablation)BLIP · COCO retrievalTR@1: 70.8 -> 73.3Improved
Replace Faster R-CNN tags with CLIP-similarity tags (ablation)BLIP · COCO image-to-text retrievalR@1: 73.3 -> 72.9No clear effect
Use the smaller ResNet101 Faster R-CNN for object tags (ablation)BLIP · COCO image-to-text retrievalR@1: 73.3 -> 72.7Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 2 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