AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

Transferable Representation Learning in Vision-and-Language Navigation

Haoshuo Huang, Vihan Jain, Harsh Mehta, Alexander Ku, Gabriel Magalhaes, Jason Baldridge, Eugene Ie

ICCV 2019 · Vision-language and multimodal › Robot action and navigation · approach: Reasoning-enhanced vision-language navigation · 88 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 CMA+NVS pretrained encoders (ablation)RCM agent · R2R Val UnseenSPL: 32.7 -> 43.0Improved
add CMA+NVS pretrained encoders (ablation)RCM agent · R2R Val UnseenSR: 41.1 -> 46.1Improved
add CMA+NVS pretrained encoders (ablation)Speaker-Follower agent · R2R Val UnseenSPL: 29.0 -> 31.2Improved
add CMA+NVS pretrained encoders (ablation)Speaker-Follower agent · R2R Val UnseenSR: 36.0 -> 40.0Improved
add CMA-only pretraining (ablation)RCM agent · R2R Val UnseenSPL: 32.7 -> 42.0Improved
add CMA-only pretraining (ablation)Speaker-Follower agent · R2R Val UnseenSPL: 29.0 -> 26.8Hurt
add NVS-only pretraining (ablation)RCM agent · R2R Val UnseenSPL: 32.7 -> 20.6Hurt
add NVS-only pretraining (ablation)Speaker-Follower agent · R2R Val UnseenSR: 36.0 -> 34.9Hurt
add NVS loss to CMA training (ablation)auxiliary alignment model · CMA validation (PR+RW), Val UnseenAUC: 72.0 -> 79.2Improved
add NVS loss to CMA training (ablation)auxiliary alignment model · CMA validation (PR+RW), Val SeenAUC: 82.6 -> 84.0Improved
warm-start both image and language encoders vs only image encoder (ablation)RCM agent · R2R Val UnseenSPL: 33.1 -> 43.0Improved
warm-start both image and language encoders vs only language encoder (ablation)RCM agent · R2R Val UnseenSPL: 35.8 -> 43.0Improved

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