AI Paper Atlas

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

MomaGraph: State-Aware Unified Scene Graphs with Vision-Language Models for Embodied Task Planning

Yuanchen Ju, Yongyuan Liang, Yen-Jen Wang, Gireesh Nandiraju, Yuanliang Ju, Seungjae Lee, Qiao Gu, Elvis Hsieh, Furong Huang, Koushil Sreenath

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Pretraining and data augmentation for VLN

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 Graph-then-Plan (generate scene graph before planning) (ablation)MomaGraph-R1 · MomaGraph-Bench overallaccuracy (%): 65.1 -> 71.6Improved
add Graph-then-Plan (ablation)GPT-5 · MomaGraph-Bench overallaccuracy (%): 66.5 -> 71.6Improved
add Graph-then-Plan (ablation)InstructBLIP-7B · MomaGraph-Bench overallaccuracy (%): 39.0 -> 39.5No clear effect
add Graph-then-Plan (ablation)InternVL2.5-8B · MomaGraph-Bench Tier 1accuracy (%): 53.6 -> 51.0Hurt
use unified spatial-functional edges instead of spatial-only (ablation)MomaGraph-R1 · MomaGraph-Bench overallaccuracy (%): 59.9 -> 71.6Improved
use unified edges instead of functional-only (ablation)MomaGraph-R1 · MomaGraph-Bench overallaccuracy (%): 64.9 -> 71.6Improved
use unified edges instead of spatial-only (ablation)LLaVA-Onevision · MomaGraph-Bench overallaccuracy (%): 54.0 -> 66.0Improved
train with RL (DAPO, graph reward) instead of SFT (ablation)Qwen2.5-VL-7B · MomaGraph-Bench overallaccuracy (%): 63.9 -> 71.6Improved
train with RL instead of SFT (ablation)Qwen2.5-VL-7B · BLINKaccuracy (%): 60.4 -> 63.5Improved
RL training instead of in-context graph examples only (ablation)Qwen2.5-VL-7B · MomaGraph-Bench overallaccuracy (%): 60.2 -> 71.6Improved
vary reward weights (w_a, w_f, w_l) (ablation)MomaGraph-R1 · MomaGraph-Bench overallaccuracy (%): 68.2 to 71.6No clear effect
RL-trained scene graph model on visual correspondenceQwen2.5-VL-7B · BLINKaccuracy (%): 58.7 -> 63.5Improved

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