Research map / Image generation / Text-to-image diffusion
ConceptWeaver: Weaving Disentangled Concepts with Flow
Jintao Chen, Aiming Hao, Xiaoqing Chen, Chengyu Bai, Chubin Chen, Yanxun Li, Jiahong Wu, Xiangxiang Chu, Shanghang Zhang
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 |
|---|---|---|---|
| add stage-aware optimization (train offset in peak-influence stage) vs uniform timesteps (ablation) | ConceptWeaver · single-concept generation | concept fidelity (qualitative) | Improved |
| add stage-aware guidance vs uniform guidance across all stages (ablation) | ConceptWeaver · compositional synthesis | artifacts / concept bleeding (qualitative) | Improved |
| increase concept guidance scale w_c (ablation) | CWG · concept personalization | fidelity to reference (qualitative) | Improved |
| ConceptWeaver vs DreamBooth | ConceptWeaver · Single-concept generation | GPT-4o score: 3.7 -> 3.8 | Improved |
| ConceptWeaver vs MS-Diffusion on DINO similarity | ConceptWeaver · Single-concept generation | DINO similarity: 70.69 -> 69.57 | Hurt |
| ConceptWeaver vs TokenVerse* reproduction | ConceptWeaver · Single-concept generation | CLIP image alignment: 61.75 -> 76.13 | Improved |
| ConceptWeaver vs MS-Diffusion in user study | ConceptWeaver · User study | Concept consistency: 3.75 -> 4.06 | Improved |
| text alignment ConceptWeaver vs TokenVerse* | ConceptWeaver · Single-concept generation | CLIP text alignment: 35.20 -> 35.22 | No clear effect |
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.
- MS-Diffusion — MS-Diffusion: Multi-subject Zero-shot Image Personalization with Layout Guidance (ICLR 2025) · better on 6 of 7 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
- You Can’t Have It Both Ways: Concept Entanglement Limits Diffusion Model Unlearning (NeurIPS 2026 · similarity 0.88)
- Unveiling Concept Attribution in Diffusion Models (NeurIPS 2025 · similarity 0.86)
- EraseFlow: Learning Concept Erasure Policies via GFlowNet-Driven Alignment (NeurIPS 2025 · similarity 0.85)
- The Hidden Language of Diffusion Models (ICLR 2024 · similarity 0.85)
- Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space (NeurIPS 2024 · similarity 0.85)
- Concept Reachability in Diffusion Models: Beyond Dataset Constraints (ICML 2025 · similarity 0.84)
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.
