# TRACE: Trajectory-Constrained Concept Erasure in Diffusion Models

arXiv: https://arxiv.org/abs/2505.23312 · HTML: https://mlsanity.com/paper/2505.23312/

## Key facts

- Cited by 2 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 24 of its 35 references that are in the map.
- First posted on arXiv 29 May 2025.

## Placed in

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

## Builds on (most cited first, 24 of 24)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Prompt-to-Prompt Image Editing with Cross Attention Control](https://mlsanity.com/paper/2208.01626/) (2022)
- [Erasing Concepts from Diffusion Models](https://mlsanity.com/paper/2303.07345/) (2023)
- [Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models](https://mlsanity.com/paper/2211.05105/) (2022)
- [Unified Concept Editing in Diffusion Models](https://mlsanity.com/paper/2308.14761/) (2023)
- [MACE: Mass Concept Erasure in Diffusion Models](https://mlsanity.com/paper/2403.06135/) (2024)
- [Tree-Ring Watermarks: Fingerprints for Diffusion Images that are Invisible and Robust](https://mlsanity.com/paper/2305.20030/) (2023)
- [TF-ICON: Diffusion-Based Training-Free Cross-Domain Image Composition](https://mlsanity.com/paper/2307.12493/) (2023)
- [Robust Watermarking Using Generative Priors Against Image Editing: From Benchmarking to Advances](https://mlsanity.com/paper/2410.18775/) (2024)
- [EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers](https://mlsanity.com/paper/2412.20413/) (2024)
- [Spontaneous Symmetry Breaking in Generative Diffusion Models](https://mlsanity.com/paper/2305.19693/) (2023)
- [Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Concepts](https://mlsanity.com/paper/2504.12782/) (2025)
- [PRISM: Self-Pruning Intrinsic Selection Method for Training-Free Multimodal Data Selection](https://mlsanity.com/paper/2502.12119/) (2025)
- [CoT-Kinetics: A Theoretical Modeling Assessing LRM Reasoning Process](https://mlsanity.com/paper/2505.13408/) (2025)
- [Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation](https://mlsanity.com/paper/2410.15618/) (2024)
- [ROBIN: Robust and Invisible Watermarks for Diffusion Models with Adversarial Optimization](https://mlsanity.com/paper/2411.03862/) (2024)
- [Towards Visual-Prompt Temporal Answering Grounding in Medical Instructional Video](https://mlsanity.com/paper/2203.06667/) (2022)
- [More but Correct: Generating Diversified and Entity-revised Medical Response](https://mlsanity.com/paper/2108.01266/) (2021)
- [ICH-SCNet: Intracerebral Hemorrhage Segmentation and Prognosis Classification Network Using CLIP-guided SAM mechanism](https://mlsanity.com/paper/2411.04656/) (2024)
- [ICHPro: Intracerebral Hemorrhage Prognosis Classification Via Joint-attention Fusion-based 3d Cross-modal Network](https://mlsanity.com/paper/2402.11307/) (2024)
- [Bilateral Personalized Dialogue Generation with Contrastive Learning](https://mlsanity.com/paper/2106.07857/) (2021)

## Built on by (most cited first, 2 of 2)

- [Closed-Form Concept Erasure via Double Projections](https://mlsanity.com/paper/2604.10032/) (2026)
- [HyperErase: Scale-Calibrated Hypernetwork for Multi-Concept Erasure in Text-to-Image Models](https://mlsanity.com/paper/2609.31154/) (2026)

## Abstract (arXiv)

Text-to-image diffusion models have shown unprecedented generative capability, but their ability to produce undesirable concepts (e.g.~pornographic content, sensitive identities, copyrighted styles) poses serious concerns for privacy, fairness, and safety. Concept erasure aims to remove or suppress specific concept information in a generative model. In this paper, we introduce TRACE (Trajectory-Constrained Attentional Concept Erasure), a novel method to erase targeted concepts from diffusion models while preserving overall generative quality. Our approach combines a rigorous theoretical framework, establishing formal conditions under which a concept can be provably suppressed in the diffusion process, with an effective fine-tuning procedure compatible with both conventional latent diffusion (Stable Diffusion) and emerging rectified flow models (e.g.~FLUX). We first derive a closed-form update to the model's cross-attention layers that removes hidden representations of the target concept. We then introduce a trajectory-aware finetuning objective that steers the denoising process away from the concept only in the late sampling stages, thus maintaining the model's fidelity on unrelated content. Empirically, we evaluate TRACE on multiple benchmarks used in prior concept erasure studies (object classes, celebrity faces, artistic styles, and explicit content from the I2P dataset). TRACE achieves state-of-the-art performance, outperforming recent methods such as ANT, EraseAnything, and MACE in terms of removal efficacy and output quality.