Research map / Image generation / Text-to-image diffusion
ACCORD: Alleviating Concept Coupling through Dependence Regularization for Text-to-Image Diffusion Personalization
Shizhan Liu, Hao ZHENG, Hang Yu, Jianguo Li
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 ACCORD (DDLoss+PDLoss) to CustomDiffusion | CustomDiffusion · DreamBench | CLIP-I: 62.7 -> 71.1 | Improved |
| add ACCORD to CustomDiffusion | CustomDiffusion · DreamBench | CLIP-T: 34.2 -> 33.9 | No clear effect |
| add ACCORD to DreamBooth | DreamBooth · DreamBench | CLIP-T: 30.3 -> 31.1 | Improved |
| add ACCORD to LoRA (SDXL) | LoRA (SDXL) · DreamBench | CLIP-T: 34.5 -> 35.1 | Improved |
| add ACCORD to LoRA (SDXL) on style | LoRA (SDXL) · StyleBench | Gram-D: 47193 -> 47693 (+1.1%) | Hurt |
| add ACCORD to VisualEncoder on style with tuned weights | VisualEncoder · StyleBench | Gram-D: 32176 -> 27984 (-13.0%) | Improved |
| add ACCORD to IP-Adapter | IP-Adapter · FFHQ | Face-Sim: 14.8 -> 16.4 | Improved |
| add PDLoss alongside DDLoss (ablation) | VisualEncoder (SD1.5) · DreamBench | CLIP-I: 79.8 -> 80.4 | Improved |
| add DDLoss to LoRA (FLUX) (ablation) | LoRA (FLUX) · DreamBench | CLIP-T: 33.4 -> 34.8 | Improved |
| apply ACCORD with only one reference image (ablation) | VisualEncoder · DreamBench | CLIP-T: 25.0 -> 24.7 (CLIP-I 75.9 -> 78.9) | Hurt |
| vary DDLoss weight 0.1-0.3 and PDLoss weight 0.001-0.003 (ablation) | CustomDiffusion · DreamBench | CLIP-I: 70.5-71.1 across settings | No clear effect |
| PDLoss target lowered (cos+1 target) (ablation) | VisualEncoder · DreamBench | CLIP-I: 79.1 -> 77.6 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Omnigen — OmniGen: Unified Image Generation (CVPR 2025) · better on 1 of 2 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
- How to Continually Adapt Text-to-Image Diffusion Models for Flexible Customization? (NeurIPS 2024 · similarity 0.90)
- Direct Consistency Optimization for Robust Customization of Text-to-Image Diffusion models (NeurIPS 2024 · similarity 0.89)
- Concept Weaver: Enabling Multi-Concept Fusion in Text-to-Image Models (CVPR 2024 · similarity 0.88)
- ClassDiffusion: More Aligned Personalization Tuning with Explicit Class Guidance (ICLR 2025 · similarity 0.88)
- PaRa: Personalizing Text-to-Image Diffusion via Parameter Rank Reduction (ICLR 2025 · similarity 0.88)
- The Hidden Language of Diffusion Models (ICLR 2024 · similarity 0.88)
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.
