AI Paper Atlas

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

Bridging the Visual Semantic Gap in VLN via Semantically Richer Instructions

Joaquín Ossandón, Benjamín Earle, Alvaro Soto

ECCV 2022 · Vision-language and multimodal › Robot action and navigation · approach: Reasoning-enhanced vision-language navigation · 2 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 Speaker pre-training with synthetic instructionsRegretful-Agent · R2R val unseensuccess rate: 0.43 -> 0.47Improved
add objects auxiliary task (lambda=0.5, N=2) to Speaker instruction generationRegretful-Agent pre-training · R2R val unseensuccess rate: 0.47 -> 0.51Improved
add crafted-instructions auxiliary task (beta=0.3) to SpeakerRegretful-Agent pre-training · R2R val unseensuccess rate: 0.47 -> 0.50Improved
add objects auxiliary task (lambda=0.5, N=2) to SpeakerRegretful-Agent pre-training · R2R val seensuccess rate: 0.71 -> 0.72Improved
increase objects per word N from 2 to 3 (lambda=0.5)Regretful-Agent pre-training · R2R val unseensuccess rate: 0.51 -> 0.47Hurt
decrease objects per word N from 2 to 1 (lambda=0.5)Regretful-Agent pre-training · R2R val unseensuccess rate: 0.51 -> 0.46Hurt
pre-train directly on rule-crafted instructions instead of Speaker outputRegretful-Agent · R2R val unseensuccess rate: 0.43 -> 0.46Mixed
add objects/crafted auxiliary-task pre-training instructionsHAMT · R2R val unseensuccess rate (%): 52.3 -> 52.4Improved
add Speaker pre-trainingHAMT · R2R val unseensuccess rate (%): 48.7 -> 52.3Improved
add visual features (vs blind agent) (ablation)Self-Monitoring · R2R val seensuccess rate: 0.32 -> 0.62Improved
add visual features (vs blind agent) (ablation)Regretful-Agent · R2R val unseensuccess rate: 0.36 -> 0.43Improved
add nouns+adjectives only (remove other words) from instructions (ablation)Regretful-Agent · R2R val unseensuccess rate: 0.43 -> 0.37Hurt

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