AI Paper Atlas

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

MULE: Multimodal Universal Language Embedding

Donghyun Kim, Kuniaki Saito, Kate Saenko, Stan Sclaroff, Bryan Plummer

AAAI 2020 · Vision-language and multimodal › Human and embodied modeling · approach: Vision-language pretraining · 37 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
Use a shared, visually-semantic multilingual embedding rather than separate language branchesImage-sentence retrieval · Multi30K Frenchmean recall: 52.0 -> 56.4Improved
Use a shared, visually-semantic multilingual embedding rather than separate language branchesImage-sentence retrieval · Multi30K Czechmean recall: 43.2 -> 50.3Improved
Use a shared multilingual model instead of an English-only modelImage-sentence retrieval · Multi30K Englishmean recall: 71.1 -> 68.0Hurt
Add multi-layer neighborhood constraints (ablation)MULE · Multi30K Czechmean recall: 43.9 -> 46.4Improved
Add a language classifier to neighborhood constraints (ablation)MULE · Multi30K Czechmean recall: 46.4 -> 49.5Improved
Add MULE pretraining (ablation)MULE with neighborhood constraints and language classifier · Multi30K Czechmean recall: 49.5 -> 50.3Improved
Add neighborhood constraints (ablation)MULE · MSCOCO Chinesemean recall: 54.2 -> 56.6Improved
Add a language classifier to neighborhood constraints (ablation)MULE · MSCOCO Chinesemean recall: 56.6 -> 57.9Improved
Add MULE pretraining (ablation)MULE with neighborhood constraints and language classifier · MSCOCO Chinesemean recall: 57.9 -> 58.8Improved
Add machine-translated training sentences alongside human-generated sentencesMULE EmbN Full · Multi30K Frenchmean recall: 56.4 -> 62.3Improved
Add machine-translated training sentences alongside human-generated sentencesMULE EmbN Full · Multi30K Czechmean recall: 50.3 -> 57.7Improved
Add machine-translated training sentences alongside human-generated sentencesMULE EmbN Full · MSCOCO Chinesemean recall: 58.8 -> 74.8Improved

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