AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

SAFE-EQA: Semantic-Aware Efficient Exploration for Embodied Question Answering

Yijie Tang, Ke Xia, Jiazhao Zhang, Zhinan Yu, Zhiyuan Yu, Dezun Dong, Renjiao Yi, Chenyang Zhu, Kai Xu

ECCV 2026 · Vision-language and multimodal › Visual question answering · approach: Visual reasoning diagnostics · 0 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
use semantic-aware TSP instead of distance-only TSP (ablation)SAFE-EQA · OpenEQALLM-Match SPL: 45.7 -> 53.9Improved
use semantic-aware TSP instead of distance-only TSP (ablation)SAFE-EQA · OpenEQAtokens per scene: 14376.9 -> 10585.8Improved
add frontier-triggered priority assessment (ablation)SAFE-EQA · EXPRESS-Bench long trajectories (>25 m)trajectory length (m): 39.4 -> 32.7Improved
add frontier-triggered priority assessment (ablation)SAFE-EQA · EXPRESS-Bench long trajectories (>25 m)tokens: 35921.3 -> 38753.2Hurt
SAFE-EQA vs 3D-Mem, multi-step TSP planning with hierarchical answeringSAFE-EQA · OpenEQAtokens per scene: 19488.7 -> 10585.8Improved
SAFE-EQA vs 3D-Mem, multi-step TSP planning with hierarchical answeringSAFE-EQA · EXPRESS-Benchtokens per scene: 23270.1 -> 14991.5Improved
SAFE-EQA vs 3D-Mem, multi-step TSP planning with hierarchical answeringSAFE-EQA · HM-EQAtokens per scene: 22461.7 -> 16214.1Improved
SAFE-EQA vs 3D-Mem on exploration efficiencySAFE-EQA · OpenEQALLM-Match SPL: 42.0 -> 53.9Improved
SAFE-EQA vs 3D-Mem on answering accuracySAFE-EQA · EXPRESS-BenchLLM-Match: 67.5 -> 63.3Hurt
SAFE-EQA vs 3D-Mem on answering accuracySAFE-EQA · OpenEQALLM-Match: 52.6 -> 56.4Improved
SAFE-EQA vs 3D-Mem on answering accuracySAFE-EQA · HM-EQA†LLM-Match: 59.0 -> 59.2No clear effect

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