- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
The authors study the gap between explicit chain-of-thought reasoning and latent multimodal reasoning. In this paper, the authors propose DepLatent, which uses region-wise masking to estimate visual dependencies. These signals supervise visual alignment and step-level reconstruction in the latent space. Experiments on three VLM backbones and six benchmarks show consistent gains over latent reasoning baselines, with lower inference latency than explicit CoT.
Critical strengths
-
I think the problem is important. Improving latent multimodal reasoning while preserving inference efficiency is a meaningful direction.
-
The empirical gains are quite strong across multiple backbones and benchmarks. DepLatent consistently improves over Heima and other latent reasoning baselines.
-
The method achieves substantial speedups over explicit CoT while maintaining or improving accuracy.
Critical weaknesses
-
The validation of the visual dependency signal is not sufficiently independent of its construction. Eq. 3 defines dependency through masking-induced likelihood changes, while Fig. 3(b) evaluates it using masking-induced performance drops. Although this provides some downstream evidence, it does not establish whether the identified regions correspond to semantically meaningful reasoning evidence.
-
I am not fully convinced by the claim that weak visual grounding is the main reason latent reasoning underperforms explicit CoT. The evidence mainly relies on hidden-state sensitivity under image masking, but the connection between this metric and reasoning quality is not sufficiently justified.
-
There seem to be inconsistencies in the objective formulation. Eq. 6 uses positive cosine similarity inside a loss that is minimized, and the term contains an unbound index . Eq. 7 also reconstructs segments but sums over . These issues make the actual training objective difficult to verify. I don't know whether the authors have put enough attention on the manuscript.
-
The method introduces substantial training complexity. DepLatent requires region-wise masking, dependency-map computation, an additional visual alignment module, and a reconstruction decoder. For Llama-3.2-11B, training time increases from 48h to 64h compared with Heima.
Questions for authors
Alongside the weakness, I have the following questions:
-
Could the authors provide an independent evaluation showing that high-dependency regions correspond to semantically meaningful visual evidence, rather than simply regions that cause large likelihood changes when masked?
-
Could the authors clarify how the hidden-state distributions in the JSD-based visual sensitivity analysis are constructed, and why this metric is a reliable proxy for visual grounding?
-
Is the sign in Eq. 6 correct? If the objective is minimized, why does the visual alignment term use positive cosine similarity? Moreover, where the in Eq6 comes from?
-
In Eq. 7, should the summation index run over instead of ?
-
What is the additional computational overhead of dependency-map construction, especially for larger backbones?
3 comments
W4&Q5. Training costs
Thank you for raising the concern regarding the training cost. We agree that DepLatent introduces both implementation complexity and computational overhead. As shown in Table 1, DepLatent outperforms Heima by 7.0% on Llama-3.2-11B-Vision-Instruct, 7.4% on Qwen3-VL-8B-Instruct, and 6.1% on Qwen2.5-VL-7B-Instruct. Therefore, we view this increase in training GPU-hours as a reasonable trade-off in exchange for the performance gains during inference.
New results on larger models. Due to resource limits, we have tried our best to compare our model with Heima using Qwen2.5-VL-32B-Instruct. Both of them are trained with the same training settings: using Qwen2.5-32B-Instruct as the decoder, training steps=1K. As shown in Table R2, our DepLatent can improve over Heima by 5.4%, effectively extending into larger model backbones using a modest increase in training hours.
Table R2: Training cost (GPU-hours) and performance comparison across model backbones. GPU-hours denote cumulative GPU usage across all GPUs; total GPU-hours include dependency-map construction and latent training.
| Backbone | Method | Map construction (GPU-hours) | Map storage | Latent training (GPU-hours) | Total (GPU-hours) | EMMA accuracy |
|---|---|---|---|---|---|---|
| Qwen2.5-VL-32B-Instruct | Heima | — | — | — | ~58 | 32.3 |
| Qwen2.5-VL-32B-Instruct | DepLatent | ~18 | 0.90GB | ~50 | ~68 | 37.7 |
| Llama-3.2-11B-Vision | Heima | — | — | — | ~48 | 20.4 |
| Llama-3.2-11B-Vision | DepLatent | ~14 | 0.91GB | ~50 | ~64 | 30.7 |
| Qwen2.5-VL-7B-Instruct | Heima | — | — | — | ~24 | 22.6 |
| Qwen2.5-VL-7B-Instruct | DepLatent | ~6 | 0.63GB | ~20 | ~26 | 34.3 |
| Qwen3-VL-8B-Instruct | Heima | — | — | — | ~27 | 24.1 |
| Qwen3-VL-8B-Instruct | DepLatent | ~8 | 0.86GB | ~23 | ~31 | 34.8 |
Thanks for the time, I think the answers have addressed my concerns, no more questions. I'll raise my score to 3
We sincerely thank all reviewers for their thoughtful and constructive feedback. We appreciate the recognition of the significance of our research problem (wLNR, Kavd), the value of our empirical analysis of latent visual reasoning (UZ3v, wLNR), the motivation behind our dual-factorized supervision framework (UZ3v, wLNR), and the substantial performance gains and strong inference efficiency achieved across multiple backbones (UZ3v, wLNR, Kavd).
Below we address your questions:
W1&Q1. Validation of Visual Dependency Signal
We thank the reviewer for raising this concern. As suggested, we conducted a new independent evaluation by comparing model-selected high-dependency (threshold = 0.7) regions with human-annotated reasoning-relevant regions. Using human annotations as the reference, the comparison yields an overall precision of 92.4%, recall of 87.2%, and F1 score of 89.7%, indicating that high-dependency regions capture human-recognized reasoning evidence. We also provide qualitative visualizations in Appendix B.
Table R1. Validation of visual dependency signal with human review on 100 randomly sampled data. H: human-selected regions; M: model-selected regions; C: matched regions. Precision =C/M; Recall =C/H; F1 =2C/(H+M).
W2&Q2. Weak Visual Grounding and JSD
We understand the reviewer’s concern, and our results suggest that weak visual grounding is an important factor in the performance degradation of latent reasoning in the following aspects:
1. From an information-theoretic perspective, this metric quantifies the change in normalized hidden features induced by removing visual information. A higher value suggests that the visual input contributes substantial information to the reasoning trace. 2. The visualizations in Appendix B show the qualitative interpretation. 3. We evaluate this metric on Qwen2.5 and Qwen3 across diverse domains. As shown in Figure 2(b), latent reasoning models consistently show substantially lower visual sensitivity, supporting the consistency of this finding across models and tasks.
We will carefully revise it according to your valuable comments.
W3&Q3&Q4. Training Formulas
We apologize for the notation errors. We use the cosine distance 1−cos, averaged over reasoning tokens and visual regions. Equation (7) sums to Lv. These notation corrections do not change the actual implementation.