AI Paper Atlas

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

Hypothesis Graph Refinement: Hypothesis-Driven Exploration with Cascade Error Correction for Embodied Navigation

Peixin Chen, Guoxi Zhang, Jianwei Ma, Qing Li

ECCV 2026 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN · 3 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 hypothesis module (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 67.71 -> 72.41Improved
add semantic hypothesis module (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SPL: 50.02 -> 56.22Improved
add semantic hypothesis module (ablation)HGR · A-EQALLM-Match: 48.4 -> 55.9Improved
add semantic hypothesis module (ablation)HGR · EM-EQALLM-Match: 50.5 -> 58.3Improved
add cascade correction (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 68.61 -> 72.41Improved
add cascade correction (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SPL: 51.12 -> 56.22Improved
add cascade correction (ablation)HGR · A-EQALLM-Match: 49.6 -> 55.9Improved
cascade correction instead of local delete only (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 70.85 -> 72.41Improved
add both semantic hypothesis and cascade correction (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 63.42 -> 72.41Improved
refutation threshold 0.3 instead of 0.5 (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 72.41 -> 66.2Hurt
refutation threshold 0.7 instead of 0.5 (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SR: 72.41 -> 67.8Hurt
use GPT-4o instead of heuristic co-occurrence predictor (ablation)HGR · GOAT-Bench (subset, 278 subtasks)SPL: 52.13 -> 56.22Improved

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