Research map / Graph learning / Knowledge and temporal graphs
Reflecting Process Expertise in Procedural Material Generation
Kunal Gupta, Gaurav Joshi, Yen-Ru Chen, Seemandhar Jain, Ishit Mehta, Manmohan Chandraker
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 changed | Where | Measured | Effect |
|---|---|---|---|
| retrieve process traces instead of material graphs (ablation) | MaterialApprentice · Expert study | structural edits: 10.84 -> 7.84 | Improved |
| retrieve process traces instead of material graphs (ablation) | MaterialApprentice · Expert study | total edits: 15.67 -> 13.58 | Improved |
| retrieve process traces instead of material graphs (ablation) | MaterialApprentice · Expert study | parameter tweaks: 4.83 -> 5.75 | Hurt |
| retrieve process traces instead of material graphs (ablation) | MaterialApprentice · Fully automatic generation | expert preference: 72% | Improved |
| add process-to-graph Compiler instead of direct graph generation (ablation) | MaterialApprentice · 100 prompts | execution rate: 70% -> 100% | Improved |
| use process-space retrieval and synthesis instead of Graph->Graph (ablation) | MaterialApprentice · Matched baseline initial output | artist preference: 6 -> 76 | Improved |
| use process-space retrieval and synthesis instead of Graph->Graph (ablation) | MaterialApprentice · Matched baseline target refinement | total edits: 37.9 -> 28.1 | Improved |
| use process-space retrieval and synthesis instead of Graph->Graph (ablation) | MaterialApprentice · Instructional edit | total edits: 8.0 -> 10.7 | Hurt |
| add retrieved process demonstrations (0 to 3) (ablation) | MaterialApprentice · generation quality | quality | Improved |
| use more than three retrieved demonstrations (ablation) | MaterialApprentice · generation quality | quality | No clear effect |
| reduce video corpus to 5% of tutorials (ablation) | MaterialApprentice · generation quality | quality | No clear effect |
| remove category-matching videos from corpus (ablation) | MaterialApprentice · evaluation prompts | qualitative | No clear effect |
Closest papers by meaning
- VLMaterial: Procedural Material Generation with Large Vision-Language Models (ICLR 2025 · similarity 0.86)
- MatFuse: Controllable Material Generation with Diffusion Models (CVPR 2024 · similarity 0.86)
- Why Not Use Your Textbook? Knowledge-Enhanced Procedure Planning of Instructional Videos (CVPR 2024 · similarity 0.83)
- ShapeCraft: LLM Agents for Structured, Textured and Interactive 3D Modeling (NeurIPS 2025 · similarity 0.83)
- Material Anything: Generating Materials for Any 3D Object via Diffusion (CVPR 2025 · similarity 0.82)
- Position: You Can't Manufacture a NeRF (ICML 2025 · similarity 0.82)
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.
