- Rating
- 3(+1)
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
DualHeads adapts few-step diffusion generators using two classifiers on a frozen backbone: one learns a reward surrogate, and the other encourages agreement with reference images. The method supports black-box rewards and adds no inference overhead. Results across OCR, GenEval, and preference rewards are promising.
Critical strengths
-
The approach combines black-box reward optimization and reference matching without requiring trajectory likelihoods or an online generator-score model.
-
Experiments span multiple rewards and two backbones, with held-out quality metrics providing evidence beyond improvement on the training reward.
Critical weaknesses
-
Insufficient controls to establish the contribution. The OCR reference pool is already reward-filtered, so reference matching alone could explain part of the gain. Missing controls include reference-only training, reward-only training, and freezing the reward head after warm-up. A simple alternative worth comparing is supervised fine-tuning on the same high-reward reference images. These comparisons would establish whether online dual-head training is necessary.
-
Claims need tighter qualification. The KL interpretation applies to idealized classifiers and noised, recomputed predictions; its connection to the final sampler distribution and detached training update is not established. Patchwise classification on frozen features further restricts the density-ratio interpretation. Similarly, improved preference scores do not establish diversity preservation, and the reference-pool ablation does not support a clean separation between reward and reference-head responsibilities.
Questions for authors
-
How does full DualHeads compare with reference-only training and a frozen, warmed-up reward head under identical conditions?
-
Which claims apply exactly to the implemented update, and which are idealized interpretations? What connects the recomputed-prediction objective to final generated samples?
-
What is the exact initial SD3.5-M distilled checkpoint, and how does it perform before adaptation? Can Table 2 comparisons use matched resolutions?
-
What are the total GPU-hours, peak memory, reward-query count, and reference-pool construction cost? The heads contain 92.5M–208.3M parameters and require gradients through a frozen backbone, so practical training efficiency remains unclear.
Suggested Improvements:
-
Add the component controls above and one simple reference-data fine-tuning baseline.
-
Report task reward, held-out quality, and sample diversity together across the reward-weight sweep.
-
Clarify evaluation uncertainty and distinguish prompt-level variation from variation across training runs.
-
Position the contribution as a practical simplification; the reported results do not establish consistent superiority over existing few-step methods.
Dear Reviewer,
The authors have provided their responses to the raised concerns and revisions. Please review their responses and join the discussion before the end of the author-review discussion phase (Nov 18, AoE). Thanks.
Best, AC