AI Paper Atlas

Research map / 3D vision / 3D generation

PhyScensis: Physics-Augmented LLM Agents for Complex Physical Scene Arrangement

Yian Wang, Han Yang, Minghao Guo, Xiaowen Qiu, Tsun-Hsuan Wang, Wojciech Matusik, Joshua B. Tenenbaum, Chuang Gan

ICLR 2026 · 3D vision › 3D generation · approach: Generative indoor scene synthesis · 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 the feedback system (vs binary success/failure only) (ablation)PhyScensis · Retry Timesretries: 1.69 -> 1.04Improved
add the feedback system (ablation)PhyScensis · Time-costtime: 132.29 -> 106.41Improved
add empty-region reporting to feedback (ablation)PhyScensis · Time-costtime: 126.09 -> 106.41Improved
add empty-region reporting to feedback (ablation)PhyScensis · Retry Timesretries: 1.43 -> 1.04Improved
add visual feedback to the full framework (ablation)PhyScensis · Retry Timesretries: 1.04 -> 0.95Mixed
add designed predicates and solver instead of Random placement (ablation)PhyScensis · tabletop scenesVQA Score: 0.415 -> 0.704Improved
add designed predicates and solver instead of LLM-Only placement (ablation)PhyScensis · tabletop scenesVQA Score: 0.592 -> 0.704Improved
add solver and physics instead of LLM-Only placement (ablation)PhyScensis · tabletop scenesSettle Distance: 0.154 -> 0.003Improved
use PhyScensis instead of 3D-Generalistscene generation · tabletop scenesVQA Score: 0.578 -> 0.704Improved
use PhyScensis scenes for imitation data instead of 3D-Generalistdiffusion policy · robot reaching/placingsuccess rate: 4/10 -> 9/10 reaching; 1/10 -> 3/10 placingImproved
use PhyScensis instead of LayoutVLMscene generation · LayoutVLM comparisonSettle Distance: 0.115 -> 0.003Improved
use PhyScensis instead of ClutterGenobject arrangement · 500 sampled scenesobjects placed: 3.57 -> 14.77Improved

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