- Rating
- 1
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 1
Read review
Paper summary
This paper proposes AID-VAR, a plug-in module attached to a frozen VAR backbone to deal with cross-scale error accumulation. A light Transformer injector builds a spatial guidance map from the previous-scale feature and adds it to the hidden state of the next scale. The injector is trained adversarially against a DINO-based projected discriminator in RGB space. The paper also proposes ISCS, a metric for consistency between adjacent scales, and reports FID improvement on VAR-d16/d20/d24 and on FlexVAR in the appendix. Post-hoc enhancement of a frozen backbone is a useful direction and the idea is intuitive. However, the actual mechanism looks closer to learned guidance than to error correction. Also, the source of the baseline numbers and the validity of ISCS remain questionable.
Critical strengths
-
Post-hoc enhancement of a frozen VAR backbone is a practical setting, and in principle it applies to the fast-growing VAR family, though this is not quite proven in this paper and the injector requires retraining per backbone.
-
The motivation and explanation are clear, and Fig. 1 delivers the idea well.
Critical weaknesses
-
The framing of the paper does not match the actual mechanism, which brings up a novelty concern. The paper calls it "Adversarially Injected Diagnosis", but what happens at inference is only adding a learned additive term to the hidden state (Eq. 4). The projected discriminator, hinge loss, and softmax relaxation are all existing techniques. So it is more natural to see this as another learned guidance on top of CFG. The U-shaped guidance weight curve in Tab. 2 also looks the same as a CFG scale sweep. The injector can fool the discriminator just by making the logits sharper, so the difference from lowering the sampling temperature is unclear as well. Yet there is no discussion or quantitative comparison with VAR guidance methods such as SSG [3] and IGG [4], or with the bitwise self-correction of Infinity [2], which directly handles next-scale exposure bias.
-
The source of the baseline numbers is mixed, and backbone diversity is thin as only VAR is experimented with. The reported VAR baseline FID (3.55/2.95/2.33) matches the released checkpoint numbers in the official repository. But the IS/Precision/Recall in the same row match the original paper table, which is the checkpoint giving FID 3.30/2.57/2.09. So within one row, FID and the other metrics come from different sources. Moreover, Tab. 11 (App. E.1) says these baselines are 3-seed means, but the mean FID is the same as the repository value and the mean IS is the same as the original paper value to one decimal at all three depths ( / , / , / ). A 3-seed mean cannot match an external single number three times. So I doubt whether the baseline was actually measured. With this concern in mind, the AID-VAR FID (3.24/2.54/2.08) sits at the level of the original paper (3.30/2.57/2.09), so a large part of the gain may simply be the gap between the released checkpoint and the paper checkpoint. Also, the CFG scale and top- are not stated in the paper, so it is impossible to check how the authors measured such statistics. In addition, "various backbones" in the abstract is just a depth change of the same VAR. FlexVAR in App. E.3 reports FID only and shares the same VQ codebook family, so applicability to backbones like HART [1] and Infinity [2] is not verified. Also, with A800, a d30 experiment should have been possible.
-
ISCS is the inverse of the Fréchet Distance (FD), and the weighting inflates the result. ISCS flips FD in DINO feature space, which is not essentially different from FID. In Fig. 7 the two models are almost the same at scales 0 to 7, and the difference appears only at scales 8 and 9. With , scales 8 and 9 take roughly 75% of the total weight, so most of the weighted score comes from these two scales. Taking the inverse amplifies small FD differences at fine scales, so it is unclear whether the gap in Fig. 7 reflects a real quality difference.
Questions for authors
- Were the baseline FID, IS, Precision, and Recall each measured directly, or taken from the original paper or the official repository? Why do the 3-seed means in Tab. 11 match the repository FID and the original paper IS to one decimal? What exact checkpoint, CFG scale, and top- were used, and does the AID-VAR gain hold when the baseline is reproduced with the original paper setting?
- Are the other numbers in Tab. 1 under the same condition? MAR in Tab. 1 is 2.35/227.8/0.79/0.62, which is the number without CFG in the original MAR paper. With CFG, should it not be 1.55/303.7/0.81/0.62?
- Can you compare with VAR guidance methods such as SSG [3] and IGG [4] under the same CFG setting, or show that the gain remains when CFG is replaced with IGG? Also, could you report per-scale logit entropy with and without the injector, which would help to judge whether this is simple sharpening?
- The fake set is made by free-running inference, but is CE against the GT tokens. Is the injector training teacher-forced or free-running? If teacher-forced, the injector never sees the model's own errors, so how does it learn to correct error accumulation?
- What is the wall-clock latency with and without the injector? When CFG is used, does the injector run in both the conditional and unconditional branches?
[1] HART: Efficient Visual Generation with Hybrid Autoregressive Transformer (ICLR 2025)
[2] Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis (CVPR 2025)
[3] SSG: Scaled Spatial Guidance for Multi-Scale Visual Autoregressive Generation (ICLR 2026)
[4] Rethinking Visual Autoregressive Sampling with Information-Grounding Guidance (ICML 2026)
The initial rating below is driven mainly by Critical Weakness 2 (W2). I will reconsider it once the authors clarify W2.
10 comments
Response to Reviewer J9Ki [2/5]
(b) Correction of MAR Baseline in Table 1 (Q2)
Thank you for catching this detail. The values 2.35 / 227.8 / 0.79 / 0.62 correspond to MAR-L without CFG, whereas the reported CFG-enabled entry is 1.55 / 303.7 / 0.81 / 0.62 (FID / IS / Precision / Recall).
When compiling the unified comparison table, we initially followed the reporting convention of recent AR work (such as HMAR [3]), which list the default non-CFG entry of MAR alongside CFG-enabled baselines. However, we fully agree with your suggestion that mixing CFG settings without explicit labels can be misleading. To make the comparison completely transparent, we have updated Table 1 in the revised manuscript to explicitly include and label both MAR-L (w/o CFG) (2.35 / 227.8 / 0.79 / 0.62) and MAR-L (w/ CFG) (1.55 / 303.7 / 0.81 / 0.62), along with clear CFG annotations for all baselines.The corresponding clarifications have also been incorporated into the main text of the revised manuscript.
(c) Architectural Diversity: Full FlexVAR Metrics & Bitwise Text-to-Image Infinity-2B (W2)
To verify generalizability beyond standard VAR depths and VQ codebooks, we:
- Report all evaluation metrics for FlexVAR and VAR-d30:
FlexVAR: FID 3.05, sFID 7.91, IS 282.4, Precision 0.81, Recall 0.55AID-FlexVAR-d16 (Ours): FID 2.81, sFID 7.75, IS 288.8, Precision 0.81, Recall 0.55VAR-d30: FID 1.98, sFID 8.37, IS 315.9, Precision 0.81, Recall 0.59AID-VAR-d30 (Ours): FID 1.90, sFID 8.37, IS 317.6 , Precision 0.82, Recall 0.60
- Transferred
AID-VARtoInfinity-2B[6] (Bitwise Multi-Scale Quantization & Text-to-Image):
We trained our lightweight injector on top of the frozen Infinity-2B backbone (32 layers, bitwise residual quantization tokenizer, / 0.06M pixels) using the Fine-T2I (synthetic_enhanced_prompt_random_resolution) subset, and evaluated on the standard GenEval benchmark (cfg=4.0, tau=1.0, rewrite=0):
| Model (Bitwise Text-to-Image on GenEval) | Single Obj | Two Obj | Counting | Colors | Position | Color Attr | Overall |
|---|---|---|---|---|---|---|---|
| Infinity-2B [6] (Baseline) | 80.31% | 18.19% | 7.82% | 53.44% | 3.46% | 5.74% | 28.16% |
| AID-Infinity-2B (Ours) | 84.06% | 20.96% | 9.69% | 57.45% | 7.25% | 7.00% | 31.07% |
| Gain | +3.75% | +2.77% | +1.87% | +4.01% | +3.79% | +1.26% | +2.91% (+10.3% rel.) |
AID-VAR yields consistent gains across all six GenEval compositional and structural subtasks on Infinity-2B (including more than doubling the Position score from 3.46% to 7.25%), proving that our framework generalizes effectively to bitwise tokenizers and text-to-image backbones. Notably, since Infinity-2B already incorporates bitwise self-correction during backbone training to address next-scale exposure bias, the fact that attaching our frozen-backbone AID-VAR injector on top of Infinity-2B yields an additional gain directly demonstrates that our post-hoc external guidance is orthogonal and complementary to internal self-correction mechanisms (W1).
Response to Reviewer J9Ki [3/5]
2. Comparison with SSG [4] and IGG [5], Logit Entropy, and Error Correction (W1 & Q3)
Framing Alignment and Manuscript Revision (W1)
Thank you for this insightful observation. We agree that while the discriminator diagnoses scale-level visual artifacts during training, the inference-time operation is a learned, trajectory-conditioned spatial guidance injection (Eq. 4) rather than explicit online error detection. In the revised manuscript, we have refined the Abstract and Sections 1 & 3 to align our terminology more accurately with the underlying mechanism—framing AID-VAR as discriminator-supervised spatial guidance for cross-scale error mitigation—while clarifying how it differs from training-free logit guidance (CFG/SSG/IGG) and internal bitwise self-correction (Infinity). The corresponding explanations and clarifications have also been incorporated into the revised main text.
(a) Quantitative Comparison & Complementarity with SSG [4] and IGG [5] (Q3)
We evaluated AID-VAR against and in combination with SSG [4] (cfg=1.5, ssg=1.0) and IGG [5] (replacing CFG with pure IGG, igg=1.0) on VAR-d16:
Guidance Method (VAR-d16, 50K samples) | FID | sFID | IS | Precision | Recall |
|---|---|---|---|---|---|
| VAR + CFG (1.5) | 3.61 | 8.28 | 271.7 | 0.84 | 0.51 |
| VAR + CFG (1.5) + SSG [4] | 3.29 | 7.99 | 287.2 | 0.81 | 0.53 |
| AID-VAR + CFG (1.5) (Ours) | 3.24 | 7.85 | 280.0 | 0.85 | 0.51 |
| AID-VAR + CFG (1.5) + SSG [4] (Combined) | 3.12 | 7.73 | 292.4 | 0.84 | 0.51 |
| VAR + Pure IGG [5] (Replacing CFG) | 4.40 | 8.50 | 250.6 | 0.79 | 0.54 |
| AID-VAR + Pure IGG [5] (Replacing CFG) | 4.01 | 8.30 | 257.9 | 0.80 | 0.53 |
- Competitive Standalone Performance: Under the same
cfg=1.5setting,AID-VAR(3.24FID) outperformsVAR + SSG(3.29FID) without requiring extra forward passes through the heavy VAR backbone. - Complementarity to
SSGandIGG: CombiningAID-VARwithSSGfurther reduces FID from3.29to3.12(and sFID from7.99to7.73). Similarly, when CFG is replaced byIGG, addingAID-VARreduces FID from4.40to4.01(+7.3 IS). Unlike training-free logit extrapolation (CFG/SSG/IGG),AID-VARinjects learned spatial feature corrections informed by an image-space discriminator, supporting compatibility with the tested sampling-time guidance settings. These FID improvements accompany Recall decreases from 0.533 to 0.506 with SSG and from 0.538 to 0.526 with IGG.
(b) Per-Scale Logit Entropy Analysis: Is AID-VAR Just Logit Sharpening? (W1 & Q3)
To test whether the injector merely sharpens logits (equivalent to lowering sampling temperature), we measured the per-scale token distribution entropy across 50,000 samples on VAR-d16 (cfg=1.5, top-k=900, top-p=0.96):
| Scale (Patch Grid) | 0 () | 1 () | 2 () | 3 () | 4 () | 5 () | 6 () | 7 () | 8 () | 9 () |
|---|---|---|---|---|---|---|---|---|---|---|
| VAR Entropy | 4.6437 | 5.9384 | 6.4199 | 6.8376 | 6.6736 | 6.8485 | 7.0079 | 6.5988 | 6.3072 | 5.6758 |
| AID-VAR Entropy | 4.6421 | 5.9559 | 6.4615 | 6.8861 | 6.7187 | 6.8879 | 7.0348 | 6.6346 | 6.3387 | 5.6921 |
| Relative Change | -0.03% | +0.29% | +0.65% | +0.71% | +0.68% | +0.58% | +0.39% | +0.54% | +0.50% | +0.29% |
Lowering temperature reduces entropy across scales. In contrast, AID-VAR does not reduce logit entropy—in fact, entropy slightly increases (+0.29% to +0.71%) across Scales 1–9. Furthermore, in our matched-perturbation experiment (injecting Gaussian noise at Scale 5; see Response to Reviewer #1, Section 1), AID-VAR dampens trajectory divergence (LPIPS to clean trajectory) to 0.0005 at Scale 6 (vs. 0.3208 for VAR) and 0.3160 at Scale 10 (vs. 0.5625 for VAR), providing separate evidence of reduced perturbation sensitivity. Neither scale-averaged entropy nor this single intervention substitutes for a temperature-matched quality comparison.
Response to Reviewer J9Ki [1/5]
We thank Reviewer J9Ki for the thorough, insightful review and for explicitly stating a willingness to reconsider the rating once Critical Weakness 2 (W2) is clarified. We address W2 (Q1 & Q2) first with full transparency and unified re-evaluations, followed by W1 (Q3), Q4, W3, and Q5.
1. Clarification & Unified Re-Evaluation of Baselines and New Backbones (W2, Q1 & Q2)
(a) Source of Original Baseline Numbers & Unified 3-Seed Re-Evaluation (Q1)
We sincerely apologize for the baseline reporting errors in Tables 1 and 11, and thank the reviewer for identifying them. We distinguish the two issues explicitly.
To ensure complete rigor and apple-to-apple fairness, we have replaced all baseline numbers in Tables 11 with our directly measured 3-seed results using the exact same official checkpoints and sampling settings (cfg=1.5, top-k=900, top-p=0.96, 50,000 images × 3 seeds:
seed 0,seed 1,seed 2), and we also mark the original VAR paper's reported numbers separately with a dagger (†) for transparent reference (the scaling results onVAR-d30 (2.0B)are reported separately in Section 1(c) below):