AI Paper Atlas

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

Seeing Out of the Box: End-to-End Pre-Training for Vision-Language Representation Learning

Zhicheng Huang, Zhaoyang Zeng, Yupan Huang, Bei Liu, Dongmei Fu, Jianlong Fu

CVPR 2021 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 217 citations (OpenAlex)

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
use end-to-end SOHO with a trainable CNN and visual dictionary instead of region-based VLPTSOHO · MSCOCO 5K test* (TR)R@1: 64.4 -> 66.4Improved
use end-to-end SOHO with a trainable CNN and visual dictionary instead of region-based VLPTSOHO · VQA2.0 test-stdVQA-score: 72.91 -> 73.47Improved
use end-to-end SOHO with a trainable CNN and visual dictionary instead of region-based VLPTSOHO · NLVR² test-Paccuracy: 75.80 -> 77.32Improved
use end-to-end SOHO with a trainable CNN and visual dictionary instead of region-based VLPTSOHO · SNLI-VE testaccuracy: 78.28 -> 84.95Improved
add the visual dictionary (VD) (ablation)SOHO · MSCOCO 1K test (TR)R@1: 72.80 -> 75.50Improved
add the visual dictionary (VD) (ablation)SOHO · NLVR² test-Paccuracy: 62.61 -> 65.32Improved
set visual dictionary size to 1024 (ablation)SOHO · MSCOCO 1K test (TR)R@1: 72.80 -> 73.40Improved
set visual dictionary size to 2048 (ablation)SOHO · MSCOCO 1K test (TR)R@1: 72.80 -> 75.50Improved
set visual dictionary size to 4096 (ablation)SOHO · MSCOCO 1K test (TR)R@1: 72.80 -> 71.20Hurt
set visual dictionary size to 8192 (ablation)SOHO · MSCOCO 1K test (TR)R@1: 72.80 -> 72.10Hurt
replace the region-based BUTD inference pipeline with SOHO's whole-image CNN pipelineSOHO · V100 inference experimentinference time: 464ms -> 44msImproved

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