# Rectify Then Diffuse: Disentangling Concepts Before Denoising Trajectory Unfolds

arXiv: https://arxiv.org/abs/2608.03135 · HTML: https://mlsanity.com/paper/2608.03135/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models.
- Builds on 26 of its 47 references that are in the map.
- First posted on arXiv 4 Aug 2026.

## Placed in

- Diffusion models / [Robustness, privacy and adversarial aspects](https://mlsanity.com/diffusion/robustness-privacy-and-adversarial-aspects/)

## Builds on (most cited first, 26 of 26)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Denoising Diffusion Implicit Models](https://mlsanity.com/paper/2010.02502/) (2020)
- [Hierarchical Text-Conditional Image Generation with CLIP Latents](https://mlsanity.com/paper/2204.06125/) (2022)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Scaling Rectified Flow Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2403.03206/) (2024)
- [GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models](https://mlsanity.com/paper/2112.10741/) (2021)
- [Prompt-to-Prompt Image Editing with Cross Attention Control](https://mlsanity.com/paper/2208.01626/) (2022)
- [GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignment](https://mlsanity.com/paper/2310.11513/) (2023)
- [Stanza: A Python Natural Language Processing Toolkit for Many Human Languages](https://mlsanity.com/paper/2003.07082/) (2020)
- [Compositional Visual Generation with Composable Diffusion Models](https://mlsanity.com/paper/2206.01714/) (2022)
- [TIFA: Accurate and Interpretable Text-to-Image Faithfulness Evaluation with Question Answering](https://mlsanity.com/paper/2303.11897/) (2023)
- [Training-Free Layout Control with Cross-Attention Guidance](https://mlsanity.com/paper/2304.03373/) (2023)
- [Diffusion Self-Guidance for Controllable Image Generation](https://mlsanity.com/paper/2306.00986/) (2023)
- [BoxDiff: Text-to-Image Synthesis with Training-Free Box-Constrained Diffusion](https://mlsanity.com/paper/2307.10816/) (2023)
- [What the DAAM: Interpreting Stable Diffusion Using Cross Attention](https://mlsanity.com/paper/2210.04885/) (2022)
- [Dense Text-to-Image Generation with Attention Modulation](https://mlsanity.com/paper/2308.12964/) (2023)
- [Improving Sample Quality of Diffusion Models Using Self-Attention Guidance](https://mlsanity.com/paper/2210.00939/) (2022)
- [Compositional Text-to-Image Synthesis with Attention Map Control of Diffusion Models](https://mlsanity.com/paper/2305.13921/) (2023)
- [CONFORM: Contrast is All You Need For High-Fidelity Text-to-Image Diffusion Models](https://mlsanity.com/paper/2312.06059/) (2023)
- [Attention Calibration for Disentangled Text-to-Image Personalization](https://mlsanity.com/paper/2403.18551/) (2024)
- [Rare-to-Frequent: Unlocking Compositional Generation Power of Diffusion Models on Rare Concepts with LLM Guidance](https://mlsanity.com/paper/2410.22376/) (2024)
- [Magnet: We Never Know How Text-to-Image Diffusion Models Work, Until We Learn How Vision-Language Models Function](https://mlsanity.com/paper/2409.19967/) (2024)
- [TweedieMix: Improving Multi-Concept Fusion for Diffusion-based Image/Video Generation](https://mlsanity.com/paper/2410.05591/) (2024)
- [VSC: Visual Search Compositional Text-to-Image Diffusion Model](https://mlsanity.com/paper/2505.01104/) (2025)
- [AttenCraft: Attention-guided Disentanglement of Multiple Concepts for Text-to-Image Customization](https://mlsanity.com/paper/2405.17965/) (2024)
- [Steer Away From Mode Collisions: Improving Composition In Diffusion Models](https://mlsanity.com/paper/2509.25940/) (2025)

## Built on by (most cited first, 1 of 1)

- [ToPO: Token-Conditioned Preference Routing for Attention-Based Latent Diffusion Models](https://mlsanity.com/paper/2609.03688/) (2026)

## Abstract (arXiv)

Text-to-image diffusion models can generate individual concepts well, but they often omit or merge concepts incorrectly with multiple concepts. We trace these failures to an early coordination bottleneck: before denoising begins, prompt-conditioned attention may allocate different concepts to strongly overlapping spatial support, which can keep their attention coupled as denoising proceeds. This observation motivates treating compositional generation as a boundary-condition problem rather than repeatedly controlling the evolving trajectory. To this end, we propose Rectify-then-Diffuse (RTD), a training-free framework that rectifies the initial allocation once before standard denoising. Firstly, we propose Soft-Overlap Disentanglement (SOD), which converts normalized overlap between pilot concept maps into a differentiable and layout-agnostic separation objective. Secondly, we introduce Isotropic Gradient Rectification (IGR), which normalizes the SOD gradient and applies a bounded latent displacement with a consistent scale across prompts and initializations. Extensive experiments show that RTD achieves state-of-the-art compositional fidelity and robust gains. On the AE-Bench object pair subset, RTD improves BLIP-VQA by 45.8% and ImageReward by 19.6% over CO3 while running 2.3× faster. Code will be released at https://github.com/Z-yiwei/rectify-then-diffuse