AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Multi-Lingual Acquisition on Multimodal Pre-training for Cross-modal Retrieval

Liang Zhang, Anwen Hu, Qin Jin

NeurIPS 2022 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 0 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
remove Language Exposure stage (ablation)MLA-CLIP · MSRVTT zero-shot meantext-to-video R@1: 16.7 -> 17.8Improved
add LE stage after NLT (stage two) (ablation)MLA-CLIP · Multi30K deaverage recall: 76.3 -> 78.7Improved
do LE at stage one instead of NLT (ablation)MLA-CLIP · Multi30K deaverage recall: 76.3 -> 68.2Hurt
joint NLT+LE at stage one (ablation)MLA-CLIP · Multi30K deaverage recall: 76.3 -> 71.1Hurt
optimize NLT and LE together at stage two (ablation)MLA-CLIP · Multi30K deaverage recall: 78.7 vs 78.4No clear effect
remove language acquirers (ablation)MLA-CLIP · Multi30K csaverage recall: 65.7 -> 70.8Improved
remove M-BERT embedding initialization (ablation)MLA-CLIP · Multi30K csaverage recall: 69.4 -> 70.8Improved
compare with translate-then-retrieve (CLIP+MT)MLA-CLIP · Multi30K deaverage recall: 75.3 -> 78.7Improved
use CLIP ViT-B/16 instead of ViT-B/32 backboneMLA-CLIP · Multi30K de zero-shotaverage recall: 78.7 -> 80.8Improved
fine-tune only language acquirers on downstream data (FT-All) after FT-EnMLA-CLIP · Multi30K deaverage recall: 82.6 -> 86.8Improved
add text-text plus image-text data for Korean low-resource (600 samples)MLA-CLIP · XTD Koreanaverage recall: 78.7 -> 80.1Improved
train on unseen languages after seen onesUC2 · XTD enaverage recall: 71.8 -> 63.6Hurt

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