AI Paper Atlas

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

Uncertainty-Aware Gaussian Map for Vision-Language Navigation

Jianzhe Gao, Rui Liu, Yuxuan Xu, Tongtong Cao, Yingxue Zhang, Zhanguang Zhang, Sida Peng, Yi Yang, Wenguan Wang

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 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 Semantic Gaussian Map (SGM) as 3D scene representation (ablation)DUET-based agent · REVERIE val unseenRGS: 32.15 -> 35.48Improved
add uncertainty values only (no raw Gaussians) (ablation)DUET-based agent · R2R val unseenSR: 72.22 -> 74.20Improved
add 3D Value Map uncertainty on top of SGM (ablation)SGM agent · R2R val unseenSR: 76.21 -> 78.32Improved
add geometric and semantic uncertainty to SGM (ablation)SGM agent · REVERIE val unseenRGS: 35.48 -> 36.96Improved
add appearance uncertainty alone to SGM (ablation)SGM agent · REVERIE val unseenRGS: 35.48 -> 35.68Improved
add all three uncertainties vs appearance alone (ablation)SGM agent · REVERIE val unseenRGS: 35.68 -> 37.65Improved
slightly prune low-scale and low-opacity Gaussians (tau_e 0.01, tau_alpha 0.002) (ablation)SGM agent · R2R val unseenSR: 77.30 -> 77.87Improved
moderate pruning (tau_e 0.015, tau_alpha 0.005) (ablation)SGM agent · R2R val unseen SR and FPSSR / FPS: SR 77.30 -> 78.32; FPS 11.2 -> 15.5Improved
aggressive pruning (tau_e 0.02, tau_alpha 0.010) (ablation)SGM agent · R2R val unseenSR: 78.32 -> 74.80Hurt
use full agent instead of VER on R2R test unseenOurs · R2R test unseenSR: 76 vs 76No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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