- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes VERA, a reinforcement learning framework that supervises an explicit, task-relevant visual description in addition to the final answer. Two separately prompted judges assess whether the description is supported by the image and whether the question and description allow an image-blind model to recover the reference answer. These binary rewards are combined with answer correctness and format compliance for DAPO training:
=R_{\mathrm{ans},i} +0.5R_{\mathrm{suff},i} +0.5R_{\mathrm{faith},i} +0.5R_{\mathrm{fmt},i}.$$ The description is generated after the reasoning trace and serves as an auxiliary supervision target. Experiments cover five backbones and eight benchmarks, reporting an average accuracy improvement of approximately 5 percentage points over the starting checkpoints and reduced output-token consumption.Critical strengths
-
VERA provides a clear and implementable formulation of two complementary requirements for task-relevant visual descriptions: support from the image and sufficiency for answering the question. Section 3 explicitly distinguishes these properties, and the separate judge interfaces turn them into practical auxiliary training signals.
-
The reported evaluation covers five backbones and eight benchmarks, with improvements in average accuracy and reductions in output-token consumption. The InternVL3.5-8B result is particularly encouraging. Table 2 also provides meaningful controlled evidence on Qwen3-VL-4B: retaining the response structure while removing both evidence rewards yields an average of approximately , compared with for full VERA. Figure 4 additionally characterizes how the evidence rewards change group-relative credit assignment on InternVL3.5-8B.
Critical weaknesses
-
The incremental contribution over closely related evidence-supervision methods is limited and lacks a direct empirical comparison. Vision-SR1 already decomposes visual perception and language reasoning and rewards visual descriptions through image-blind answer recovery. VEPA explicitly trains question-conditioned evidence using a blind-reader sufficiency reward and answer-leakage checks. CCCaption jointly optimizes informational completeness and image-grounded correctness. I recognize the distinction established in Section 3: VERA requires sufficiency for an individual question, whereas CCCaption targets the completeness of a general image description. Nevertheless, task-conditioned sufficiency is itself already central to VEPA. VERA's remaining contribution appears primarily to combine this type of utility reward with direct image-faithfulness supervision in a joint answer-generation objective. The paper does not compare against closely related reasoning methods such as Vision-SR1 or PRCO. The component ablations support the usefulness of the proposed combination in its own configuration, but do not establish whether this incremental design offers an advantage over the closest alternatives.
-
The strongest reported improvement lacks a matched training control, limiting attribution of the empirical contribution. InternVL3.5-8B improves from approximately to , the largest backbone-level gain in Table 1. The paper does analyze this backbone: Figure 3 reports training dynamics, and Figure 4 recomputes advantages after adding or removing reward components while holding sampled responses fixed. These analyses show how the rewards affect optimization signals, but do not measure the held-out accuracy of a policy retrained without the evidence rewards. The actual reward-removal training ablations in Table 2 are limited to Qwen3-VL-4B, whose gain over its initial checkpoint is approximately 3.91 points. Moreover, in the Qwen2.5-VL-based 7B comparison, Vision-R1 + VERA averages , below VL-Rethinker at and ReVisual-R1 at . Thus, the results demonstrate improvements over the starting checkpoints, but have yet to establish a clear advantage over the closest methods or isolate the source of the strongest gain. This is consequential given the limited methodological novelty.
Questions for authors
-
Section 3 already makes the distinction between visual faithfulness and task-specific sufficiency clear. Given that Vision-SR1 and VEPA already use image-blind answer recovery to supervise visual descriptions, could you substantiate the additional benefit of VERA's combined objective through a matched comparison against a closely related reasoning method, such as Vision-SR1 or PRCO? A comparison using the same initial checkpoint, training data, rollout/update budget, and evaluation protocol would help establish the empirical value of the incremental design.
-
Figure 4 provides an analysis of reward-induced advantage changes on fixed InternVL3.5-8B rollouts, while Table 2 provides retraining ablations on Qwen3-VL-4B. Could you extend the latter type of control to InternVL3.5-8B: retrain with answer-correctness and format rewards only, retaining the same response structure, data filtering, rollout/update budget, and answer-verification procedure? How much of the approximately 11.44-point improvement is retained relative to this matched control?
2 comments
Attribution of Performance Gains on InternVL3.5-8B (W2 / Q2)
2. Attribution of Performance Gains on InternVL3.5-8B (W2 / Q2)
The original Figure 4 analyzes how different reward components affect group-relative advantages on fixed rollouts. To directly measure their effects on the trained policy, we conducted additional reward-removal training ablations on InternVL3.5-8B.
All configurations use the same training data, structured output format, training budget, and evaluation protocol, with only the corresponding reward components modified.
In particular, the w/o Evidence configuration retains the think → caption → answer structure and uses only answer-correctness and format rewards. This provides a matched training control for evaluating the additional contribution of visual-faithfulness and task-sufficiency rewards.
Table B. Reward ablations on InternVL3.5-8B. Each entry reports Avg@3 accuracy (%) / average output tokens.
| Dataset | Original | VERA | w/o Rans | w/o Rfmt | w/o Evidence | w/o Rsuff | w/o Rfaith |
|---|---|---|---|---|---|---|---|
| MMStar | 70.7 / 2148 | 80.6 / 1467 | 65.2 / 2312 | 78.2 / 1996 | 77.6 / 2336 | 79.7 / 2455 | 78.4 / 2992 |
| ChartQA | 90.0 / 739 | 92.2 / 735 | 84.0 / 832 | 91.5 / 812 | 90.8 / 978 | 91.3 / 921 | 90.3 / 1121 |
| WeMath | 78.4 / 3897 | 88.5 / 3592 | 76.1 / 4213 | 80.7 / 4029 | 78.5 / 4012 | 79.9 / 3212 | 79.3 / 4433 |
| MMMath | 77.5 / 7287 | 87.8 / 5598 | 77.3 / 8221 | 79.3 / 8923 | 83.2 / 8776 | 79.5 / 7985 | 78.7 / 8221 |
| LogicVista | 57.3 / 8231 | 80.4 / 4355 | 51.5 / 8996 | 75.4 / 9023 | 67.5 / 7334 | 72.5 / 8832 | 62.3 / 6736 |
| MathVerse | 71.2 / 4725 | 86.4 / 4595 | 65.9 / 4922 | 79.6 / 4673 | 74.9 / 5442 | 84.5 / 5774 | 74.0 / 5509 |
| MathVision | 29.6 / 12004 | 49.0 / 6749 | 28.5 / 8297 | 39.1 / 9993 | 38.3 / 7663 | 38.4 / 9322 | 35.3 / 9982 |
| MathVista | 83.6 / 2554 | 84.9 / 2444 | 74.6 / 3328 | 84.1 / 3424 | 80.6 / 3311 | 82.6 / 3881 | 83.1 / 4410 |
| Avg | 69.79 / 5198 | 81.23 / 3692 | 65.39 / 5140 | 75.99 / 5359 | 73.93 / 4982 | 76.05 / 5298 | 72.68 / 5426 |
The complete VERA configuration achieves an average accuracy of 81.23%, improving upon the original InternVL3.5-8B checkpoint's 69.79% by 11.44 percentage points.
When both evidence rewards are removed, the w/o Evidence configuration still achieves 73.93%, corresponding to an improvement of approximately 4.14 percentage points over the original checkpoint. This shows that answer-correctness and format rewards alone also contribute to performance improvements.
Under the same output structure and training configuration, introducing visual-faithfulness and task-sufficiency rewards further increases accuracy from 73.93% to 81.23%, yielding an additional improvement of 7.30 percentage points. This difference corresponds to approximately 63.8% of the total 11.44-point improvement over the original checkpoint.
The remaining reward ablations further support the contribution of the complete objective. Removing answer correctness, format compliance, task sufficiency, or visual faithfulness reduces average accuracy to 65.39%, 75.99%, 76.05%, and 72.68%, respectively, all below the complete VERA configuration.
These retraining experiments complement the analysis in the original Figure 4. Figure 4 characterizes how evidence rewards modify group-relative credit assignment on fixed rollouts, while the new Table B directly measures their contribution to held-out accuracy after policy optimization.
In particular, the 7.30-percentage-point improvement over the matched answer-and-format-only control provides direct empirical evidence that the evidence-level rewards contribute substantially to the performance gains observed on InternVL3.5-8B.
We thank the reviewer for highlighting these experimental concerns. We have added direct comparisons with closely related methods and comprehensive reward ablations on InternVL3.5-8B, and will incorporate the corresponding results and analyses into the revised manuscript. We hope these additional experiments clarify the empirical contribution of VERA and address the reviewer's concerns. We would appreciate the reviewer reconsidering the assessment in light of these new results.
Methodological Contribution and Comparison with Closely Related Methods (W1 / Q1)
We thank the reviewer for the constructive comments. To address the concerns regarding comparisons with closely related methods and the attribution of performance gains on InternVL3.5-8B, we conducted two additional experiments, presented in Tables A and B.
1. Methodological Contribution and Comparison with Closely Related Methods (W1 / Q1)
We acknowledge that image-blind answer recovery has been explored in prior work, including Vision-SR1 and VEPA, as a means of evaluating the task sufficiency of visual evidence. VERA builds on this direction by jointly incorporating visual faithfulness and task sufficiency as two independent reward signals into reinforcement learning for final-answer generation.
Specifically, Vision-SR1 primarily evaluates visual descriptions through image-blind answer recovery, while VERA additionally employs an independent visual judge to verify whether the factual claims in the caption are supported by the original image. VEPA focuses on improving visual evidence representations through evidence pre-alignment, while PRCO adopts an Observer–Solver framework to optimize visual evidence through downstream reasoning feedback.
VERA integrates answer correctness, visual faithfulness, and task sufficiency within a single
think → caption → answertrajectory, without requiring an additional Observer–Solver training architecture. The two evidence judges receive different inputs and evaluate complementary properties of the generated caption.To further examine the empirical benefit of this design, we conducted additional direct comparisons with Vision-SR1 and PRCO using Qwen3-VL-4B-Thinking. The results are presented below.
Table A. Comparison with closely related methods on Qwen3-VL-4B-Thinking. Each entry reports Avg@3 accuracy (%) / average output tokens.
VERA achieves an average accuracy of 75.64% across the eight benchmarks, outperforming Vision-SR1 and PRCO by 1.58 and 1.46 percentage points, respectively. Compared with Vision-SR1, VERA achieves higher accuracy on six of the eight benchmarks. Compared with PRCO, VERA achieves higher accuracy on all eight benchmarks.
VERA also produces substantially shorter outputs, averaging 3,059 tokens, compared with 7,561 tokens for Vision-SR1 and 6,045 tokens for PRCO.
Furthermore, under the same structured caption-generation protocol, removing the visual-faithfulness and task-sufficiency rewards reduces average accuracy from 75.64% to 69.93%, a decrease of 5.71 percentage points. Since both configurations retain the same
think → caption → answerstructure, this comparison provides direct evidence of the additional benefit of evidence-level supervision. These results support the effectiveness of VERA's joint reward design and demonstrate its empirical advantages over closely related methods.