- Rating
- 3
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
Diffusion-based image editing typically first inverts a source image into noise space and then regenerates it. Inversion errors can cause the reconstructed image to deviate from the original, compromising preservation outside the edited region. Conventional CFG uses a fixed scalar and cannot readily accommodate differences in error across spatial locations and feature channels. This paper proposes POLARIS, a training-free method for tensor-level adaptive CFG. It establishes an analytical relationship between inversion error and CFG coefficients and derives directly solvable, coordinate-wise optimal guidance coefficients, enabling improved inversion accuracy through low-cost inference-time control.
Critical strengths
The method is lightweight and relatively novel, the derivation is reasonable, and the experiments show good visual results.
Critical weaknesses
The executable algorithm and the evidence connecting configurations across experiments remain insufficiently clear. The closed-form solution involves adjacent predictions, whereas the implementation description only states that predictions are “updated from the cache and replayed.” The paper needs to specify which step produces each prediction, which transition uses the resulting coefficients, whether there is a one-step lag, and all stabilization hyperparameters.
Questions for authors
Which differences in data, scheduling, or initialization account for the discrepancy between the correct replay results in Table 12 and the 50-step COCO results in Table 1? Please provide the exact sample counts, selection rules, backbone/code versions, and a configuration mapping table for all experiments.
3 comments
2. Clarification of the discrepancy between the main reconstruction and replay-strategy ablation results
The two experiments use different COCO2017 evaluation subsets:
- Table 1 (Main reconstruction experiment): COCO2017 Sub-1000, containing 1,000 images, with an MSE of 349 at 50 steps.
- Table 18 (Replay-strategy ablation): COCO2017 Sub-100, containing 100 images, with an MSE of 498 for reverse-order replay at 50 steps.
Because these values are computed over different evaluation subsets, the absolute MSE values 349 and 498 should not be directly compared as if they were measured on the same image pool. The difference does not indicate an inconsistency in the reconstruction procedure.
The replay-strategy ablation is specifically designed to investigate the effect of aligning the stored guidance coefficients with their corresponding reconstruction timesteps. Within the Sub-100 evaluation setting, reverse-order replay achieves an MSE of 498, outperforming forward-order replay (642), cosine replay (833), and fixed guidance with during reconstruction/sampling (1171).
We also clarify that the fixed- row in Table 18 refers to using fixed guidance during reconstruction/sampling after the adaptive inversion trajectory has been obtained. It is not the same protocol as using fixed guidance throughout both inversion and reconstruction, which is evaluated separately in Table 17.
We have corrected the table cross-reference in the main text: the replay-strategy ablation is presented as Table 18 in Appendix B.9, whereas Table 12 reports fallback and clipping statistics in Appendix B.2.
3. Clarification of experimental configurations
To improve reproducibility, we further summarize the experimental configurations and their correspondence to the tables in the manuscript. In particular, we distinguish the main reconstruction evaluation, timestep-wise diagnostics, numerical-stability analysis, guidance-granularity ablation, and replay-strategy ablation.
| Experiment | Backbone | Dataset and sample count | Scheduler / Steps | Guidance configuration |
|---|---|---|---|---|
| Main reconstruction (Table 1) | SD1.5 | COCO2017 Sub-1000 (1,000 images); Pick-a-Pic (sample count not separately specified in the table) | Deterministic DDIM, 10–100 steps | Fixed CFG 7.5 vs. tensor-valued POLARIS; reverse-order replay |
| Timestep-wise reconstruction diagnostic (Table 2) | SD1.5 | Pick-a-Pic (sample count not separately specified in the table) | 50-step DDIM | Fixed vs. POLARIS; predicted clean-sample MSE tracked across reconstruction steps |
| Local cycle-residual analysis (Table 9) | SD1.5 | COCO2017, 100 images | 49 adjacent transitions | Fixed guidance vs. POLARIS |
| Numerical-stability analysis (Tables 10–12) | SD1.5 | Same 100-image evaluation; 49 coefficient updates per image | 49 transitions; latent dimensions | FP32 denominator direction, , previous-coefficient fallback, and clipping to |
| Guidance-granularity ablation (Table 13) | DDIM-based inversion setting | Dataset and sample count not separately restated in the table caption | 50-step DDIM | Fixed, Scalar-, Channel-, Spatial-, and Tensor-POLARIS |
| Replay-strategy ablation (Table 18) | DDIM-based reconstruction | COCO2017 Sub-100 (100 images) | 10–100 inference steps | Reverse-order, forward-order, cosine-decay, and fixed- sampling |
| FreeInv / ReNoise integration (Table 7) | SD1.5 | Pick-a-Pic, 100 samples | 50 steps | Original host pipeline vs. host pipeline + POLARIS; seed 42 |
Several additional implementation details are worth emphasizing.
First, the main reconstruction experiment uses the same SD1.5 backbone and deterministic DDIM scheduler for the fixed-scale baseline and POLARIS. The principal difference is whether the guidance coefficient remains fixed at 7.5 or is adaptively updated as a tensor across latent coordinates and timesteps.
Second, the numerical-stability statistics in Tables 10–12 are collected from 100 images over 49 coefficient-update transitions per image, yielding 4,900 image-transition pairs. The latent dimension for this analysis is . The reported fallback frequency measures coordinate-wise denominator degeneracy rather than solver failures.
Third, the guidance-granularity ablation in Table 13 uses the same 50-step DDIM inversion setting across all parameterizations, ensuring that the comparison isolates the effect of guidance granularity rather than differences in the sampling schedule.
Finally, the FreeInv and ReNoise integration experiments use a matched setting with SD1.5, 100 Pick-a-Pic samples, 50 inference steps, and seed 42. These experiments evaluate the incremental benefit of integrating POLARIS into the respective host pipelines, rather than presenting a standalone ranking of the inversion methods.
We have consolidated these settings in the revised appendix and explicitly distinguish configurations that are shared across experiments from those specific to individual evaluations. Exact checkpoint identifiers, preprocessing details, and implementation versions have been reported from the corresponding experiment configurations rather than inferred from the benchmark settings.
We thank the reviewer for helping us improve the clarity, implementation transparency, and reproducibility of the manuscript.
We thank the reviewer for the constructive feedback and for identifying the need for a clearer description of the executable algorithm and the correspondence between experimental configurations.
Below we address all raised concerns point by point.
1. Clarification of the executable algorithm and stabilization settings
We clarify the actual inversion and reconstruction procedure below.
At inversion step t, the currently available coefficient ωt is used to combine the unconditional and source-conditioned predictions at the current latent xt,
ϵ^t=ϵu,t+ωt⊙(ϵc,t−ϵu,t).
The DDIM inversion transition then uses this guided prediction to produce xt+1. After xt+1 is obtained, the unconditional and conditional predictions at this new latent are evaluated and cached. Together with the cached predictions at xt, they provide the adjacent prediction differences,
Δϵu=ϵu,t+1−ϵu,t,
Δϵc=ϵc,t+1−ϵc,t.
For each latent coordinate i, we define the direction
di=Δϵc,i−Δϵu,i,
and compute the raw coefficient as
ωraw,t+1,i=−diΔϵu,i.
The newly computed coefficient ωt+1 is not used to generate xt+1. It is first applied to the subsequent inversion transition. Thus, the implementation has an explicit one-step lag. The coefficient ωt generates xt+1, while ωt+1 is computed from the adjacent predictions at xt and xt+1 and controls the next transition. This ordering avoids circular dependence between the latent state and the coefficient estimated using the prediction at that state.
For numerical stability, the SD1.5 implementation uses the following coordinate-wise safeguards.
The coefficient sequence is initialized as ω0=s01, consistent with the implementation described in the paper.
During reconstruction, the stored inversion-order coefficient sequence is replayed in reverse order. Starting from xT, reconstruction applies ωT,ωT−1,…,ω1 to the corresponding reverse transitions. Equivalently, at reconstruction-loop index k, the applied coefficient is ωT−k+1. This ensures that each stored coefficient is paired with its corresponding noise level.
The prediction cache is also important for the computational cost. The predictions evaluated at xt+1 are retained and reused as the predictions at the next inversion step. Therefore, coefficient estimation does not require additional denoiser evaluations. The total number of denoiser evaluations remains the same as that of the corresponding CFG pipeline.