AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Physics Context Builders: A Modular Framework for Physical Reasoning in Vision-Language Models

Vahid Balazadeh, Mohammadmehdi Ataei, Hyunmin Cheong, Amir Hosein Khasahmadi, Rahul G. Krishnan

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: Spatial reasoning in VLMs · 20 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 PCB (human narration) contextGPT-4o · Falling Tower sim, tower stableaccuracy: 59.6 -> 85.1 (+25.5)Improved
add PCB (human narration) contextGPT-4o-mini · Falling Tower sim, tower stableaccuracy: 53.1 -> 84.7 (+31.6)Improved
add PCB (human narration) contextGemini 1.5 Pro · Falling Tower real, tower stableaccuracy: 60.0 -> 57.9 (-2.1)Hurt
add PCB (human narration) contextGPT-4o · Falling Tower real, tower stableaccuracy: 55.0 -> 70.0 (+15.0)Improved
add PCB (human narration) contextGemini 1.5 Pro · Falling Tower sim, object above/belowaccuracy: 97.8 -> 97.4 (-0.4)No clear effect
add PCB (human narration) contextGPT-4o-mini · CLEVRER descriptiveaccuracy: 49.5 -> 65.7 (+16.2)Improved
add PCB (human narration) contextGemini 1.5 Pro · CLEVRER explanatory per questionaccuracy: 15.7 -> 35.6 (+19.9)Improved
add PCB (human narration) contextGPT-4o-mini · CLEVRER counterfactual per questionaccuracy: 15.6 -> 17.3 (+1.7)Improved
use human narration instead of structured physics context (ablation)GPT-4o · CLEVRER explanatory per optionaccuracy: 61.1 -> 67.0Improved
use human narration instead of structured physics context (ablation)GPT-4o · CLEVRER counterfactual per questionaccuracy: 19.2 -> 28.2Improved
add descriptive QAs to stability QAs in fine-tuning (ablation)PaliGemma-3B · Falling Tower sim, obj above/belowaccuracy: 41.2 -> 100.0Improved
add stability QAs to descriptive QAs in fine-tuning (ablation)PaliGemma-3B · Falling Tower sim, tower stableaccuracy: 39.1 -> 87.6Improved

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