AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Unified Visual-Semantic Embeddings: Bridging Vision and Language With Structured Meaning Representations

Hao Wu, Jiayuan Mao, Yufeng Zhang, Yuning Jiang, Lei Li, Weiwei Sun, Wei-Ying Ma

CVPR 2019 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 166 citations (OpenAlex)

Check 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
factorize caption meaning into objects, attributes, relations, and scenes, then align semantic components contrastively with image regionsUniVSE · MS-COCO 1K test image-to-sentence retrievalR@1: 57.7 -> 64.3Improved
factorize caption meaning into objects, attributes, relations, and scenes, then align semantic components contrastivelyUniVSE · MS-COCO 5K testrsum: 299.1 -> 324.8Improved
add a bag-of-components coverage embedding to the sentence embedding (ablation)UniVSE · MS-COCO adversarial image-to-sentence retrievaltotal sum: 572.4 -> 603.0Improved
add object components to the coverage embedding (ablation)UniVSE · MS-COCO object attackR@1: 40.7 -> 42.9Improved
add attribute components to the coverage embedding (ablation)UniVSE · MS-COCO attribute attackR@1: 30.0 -> 37.4Improved
add relation components to the coverage embedding (ablation)UniVSE · MS-COCO relation attackR@1: 32.6 -> 42.8Improved
combine object, attribute, and relation components in the coverage embedding instead of using only one type (ablation)UniVSE · MS-COCO adversarial image-to-sentence retrieval, total across attackstotal sum: 603.0 vs. 589.7Improved
train with contrastive examples from all semantic component types instead of object examples only (ablation)UniVSE · MS-COCO unified text-to-image retrievalmAP sum: 164.30 -> 170.42Improved
train with contrastive examples from all semantic component types instead of attribute examples only (ablation)UniVSE · MS-COCO unified text-to-image retrievalmAP sum: 155.81 -> 170.42Improved
train with contrastive examples from all semantic component types instead of relation examples only (ablation)UniVSE · MS-COCO unified text-to-image retrievalmAP sum: 168.94 -> 170.42Improved
use image matching scores to resolve ambiguous word dependenciesUniVSE · MS-COCO attribute dependency recoveryaccuracy: 37.41 -> 64.82Improved
use image matching scores to resolve ambiguous relational dependenciesUniVSE · MS-COCO relational phrase dependency recoveryaccuracy: 31.90 -> 62.69Improved

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