- Rating
- 2
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes NuST, a concept erasure method that needs no update to the diffusion model parameters. It collects activations from feed-forward modules and denoising timesteps, builds a null space from the spatial mean of benign prompts, and then learns a target-to-anchor transform by closed-form regularized regression. At inference, the transform is applied position by position, which suppresses the target concept while keeping other generation abilities intact. The authors test nudity, artistic style, and object erasure on Stable Diffusion 1.5. They evaluate robustness and generation quality with prompt attacks, classification accuracy, FID, and CLIP metrics, and they report a low target recovery rate.
Critical strengths
- The idea is clear. It steers the target concept toward a safe anchor, while it constrains the disturbance to benign representations.
- The closed-form solution needs no backpropagation and changes no original model parameter, so deployment stays simple.
- The experiments cover three erasure tasks, and they report target suppression together with non-target preservation.
Critical weaknesses
- The key baseline numbers are inconsistent. In Table 1, ActErase detects only 6 exposed body sites, yet Table 2 reports a 1.9% success rate on I2P images. With the 4703 images stated in the paper and one shared detection protocol, the two cannot correspond.
- The preservation guarantee is narrow. The null space constraint protects only the spatial mean activation, but inference applies the transform position by position. That does not directly guarantee that the actual benign representations stay unchanged.
- The core mechanism lacks a direct ablation. There is no control that holds every other setting fixed and removes only the null space constraint, so its independent contribution is hard to confirm.
- The loss in non-target capability is clear. Mean non-target object accuracy falls from 81.36% to 69.41%, and after erasing Church it is only 54.15%. That is hard to square with the word non-destructive.
- The practical cost is not light. Generating 1000 images takes 6313 seconds, about 7.9 times that of ActErase, and it needs another 6,135 seconds of preparation. (Table 6)
Questions for authors
- Do Table 1 and Table 2 use the same I2P samples, the same detection categories, and the same denominator?
- How much do the erasure rate and the preservation rate change once the null space constraint is removed?
- If the spatial mean is unchanged, how do you guarantee that a position-wise intervention does not damage benign content?
- What are the eigenvalue threshold and the lower bound on the dimension for the adaptive null space?
- Were the adversarial prompts re-optimized against the full model that includes the NuST intervention?
- Are the results stable when the retained prompts or the random seed change?
- With another sampler or with 20 sampling steps, do all the transforms have to be rebuilt?
- The original model already has zero classification accuracy on Da Vinci. How does that test show whether erasure works?