AI Paper Atlas

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

Fine-Grained Semantically Aligned Vision-Language Pre-Training

Juncheng Li, XIN HE, Longhui Wei, Long Qian, Linchao Zhu, Lingxi Xie, Yueting Zhuang, Qi Tian, Siliang Tang

NeurIPS 2022 · Vision-language and multimodal › Human and embodied modeling · approach: Language-grounded scene text and structure · 29 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 token-level Shapley interaction loss L_TSA to dual-encoder backbone (ablation)LOUPE backbone · COCO detection (CC3M ablation)mAP@0.3: 3.8 -> 7.6Improved
add L_TSA (ablation)LOUPE backbone · MSCOCO I2T R@1 (CC3M ablation)R@1: 31.0 -> 32.4Improved
add semantics-level Shapley interaction loss L_FSA (ablation)LOUPE with L_TSA · COCO detection (CC3M ablation)mAP@0.3: 7.6 -> 9.4Improved
add L_FSA (ablation)LOUPE with L_TSA · MSCOCO T2I R@1 (CC3M ablation)R@1: 26.2 -> 28.3Improved
add uncertainty-aware neural Shapley interaction learning (UNSIL) (ablation)LOUPE · MSCOCO I2T R@1 (CC3M ablation)R@1: 33.5 -> 33.3No clear effect
add UNSIL (ablation)LOUPE · training timesec/iter: 9.90 -> 1.93Improved
add UNSIL (ablation)LOUPE · COCO detection (CC3M ablation)mAP@0.3: 9.4 -> 9.0No clear effect
add UNSIL (ablation)LOUPE · RefCOCO val (CC3M ablation)accuracy: 4.1 -> 4.5No clear effect
LOUPE vs FILIP, with less pre-training data (240M vs 340M)LOUPE · MSCOCO zero-shot T2IR@1: 45.9 -> 50.1Improved
LOUPE vs CLIP on zero-shot classificationLOUPE · 11-dataset zero-shot classificationavg top-1 accuracy: 73.0 -> 76.1Improved
LOUPE vs CLIP on FER2013 and CIFAR10LOUPE · zero-shot classificationtop-1 accuracy: 57.7 -> 53.3 (FER2013)Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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