AI Paper Atlas

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

Language-Agnostic Visual-Semantic Embeddings

Jonatas Wehrmann, Douglas M. Souza, Mauricio A. Lopes, Rodrigo C. Barros

ICCV 2019 · Vision-language and multimodal › Cross-modal learning · approach: Probabilistic embeddings for cross-modal retrieval · 54 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 the non-local module (ablation)LIWE · Flickr30k image-to-textR@1: 61.8 -> 66.4Improved
use non-linear NLM instead of linear NLM (ablation)LIWE · Flickr30k image-to-textR@1: 60.7 -> 66.4Improved
reduce input dimensionality before NLM query (ablation)LIWE · Flickr30k image-to-textR@1: 56.3 -> 66.4Improved
use proposed loss instead of loss from [19] (ablation)CLMR · Flickr30k image-to-textR@1: 60.4 -> 65.8Improved
use proposed loss instead of loss from [19] (ablation)LIWE(256,256) · Flickr30k image-to-textR@1: 63.2 -> 65.3Improved
use proposed loss instead of loss from [10] (ablation)CLMR and LIWE · Flickr30kR@1: Diverges -> 65.8Improved
concatenate pretrained Glove to character-based embeddingsLIWE · MS COCO testrecall sum (Σ): 498.6 -> 507.7Improved
concatenate pretrained Glove to character-based embeddingsLIWE · Flickr30k testrecall sum (Σ): 458.1 -> 474.3Improved
replace word embeddings with character-based LIWECLMR · Multi30k German testrecall sum (Σ): 357.3 -> 392.0Improved
replace word embeddings with character-based LIWECLMR · YJ Captions testrecall sum (Σ): 451.4 -> 448.4Hurt
replace word embeddings with character-based LIWECLMR · MS COCO testrecall sum (Σ): 500.3 -> 498.6No clear effect
use LIWE instead of BERT-MultilingualCLMR-style co-training · Multi30k English image-to-textR@1: 62.0 -> 64.4Improved

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