AI Paper Atlas

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

Learning a Recurrent Residual Fusion Network for Multimodal Matching

Yu Liu, Yanming Guo, Erwin M. Bakker, Michael S. Lew

ICCV 2017 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 163 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 RRF block to the plain matching networkbaseline network · Flickr30K test image-to-textR@1: 45.0 -> 47.6Improved
add RRF block to the plain matching networkbaseline network · Flickr30K test text-to-imageR@1: 33.6 -> 35.4Improved
increase recurrent steps from T=3 to T=4 (ablation)RRF-Net · Flickr30K test image-to-textR@1: 47.6 -> 46.2Hurt
increase recurrent steps from T=1 to T=3 (ablation)RRF-Net · Flickr30K test image-to-textR@1: 46.4 -> 47.6Improved
add fusion module to recurrent residual model (ablation)RRF-Net · Flickr30K test image-to-textR@1: 45.8 -> 47.6Improved
use convolutional fusion instead of sum fusion (ablation)RRF-Net · Flickr30K test image-to-textR@1: 47.1 -> 47.6Improved
use bi-rank loss instead of bi-directional loss (ablation)baseline network · Flickr30K test image-to-textR@1: 43.4 -> 45.0Improved
use bi-rank loss instead of bi-directional loss (ablation)RRF-Net · Flickr30K test image-to-textR@1: 46.4 -> 47.6Improved
ensemble four RRF-Net models with T=1..4RRF-Net · Flickr30K test image-to-textR@1: 47.6 -> 50.8Improved
ensemble four RRF-Net models with T=1..4RRF-Net · Flickr30K test text-to-imageR@1: 35.4 -> 37.6Improved
add RRF blocks, cross-dataset transferbaseline network · Train Flickr30K, test MSCOCO image-to-textR@1: 23.7 -> 24.8Improved
use ResNet-152 instead of VGG-19 visual featuresRRF-Net · Flickr30K test image-to-textR@1: 42.1 -> 47.6Improved

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