- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
UnI-R feeds imagined future states into visual reasoning through token selection and a query-based bridge. Two-stage training encourages consistent representations and factual-future grounding. Experiments show gains across dynamic and general benchmarks and multiple backbones.
Critical strengths
The connection between future generation and reasoning is interesting, with substantial gains across tasks and backbones. Component ablations provide useful supporting evidence.
Critical weaknesses
The paper does not clearly establish whether the trained model uses future content at inference. Without controlled removal or replacement tests, the gains cannot be cleanly attributed to the imagined futures.
Questions for authors
- What is the end-to-end inference latency and memory overhead relative to the backbone and competing methods, including future generation, ACDC, and F2R? How does accuracy trade off against this cost?
- How does ACDC compare with random selection at the same budget, and how does the same bridge perform when reading current-image features instead of future states?
5 comments
Response to Reviewer XNjd (Part 2)
Q1.
What is the end-to-end inference latency and memory overhead relative to the backbone and competing methods, including future generation, ACDC, and F2R? How does accuracy trade off against this cost?
We thank the reviewer for this question. We add an efficiency evaluation on V*, reporting end-to-end latency, component timings, peak GPU memory, and accuracy for the backbone, competing methods, and UnI-R under different future generation budgets.
Rebuttal Table 3. Inference efficiency and accuracy trade-offs with backbone and competing Methods.
| Method | Future generations (s) | ACDC (s) | F2R (s) | Latency (s) | Memory (GB) | V* |
|---|---|---|---|---|---|---|
| Cosmos3 | - | - | - | 0.85981 (1.00x) | 30.71 | 75.92 |
| +VISER | - | - | - | 0.86820 (1.01x) | 30.95 | 75.92 |
| +ICoT | - | - | - | 2.00165 (2.33x) | 31.12 | 76.44 |
| +Monet | - | - | - | 2.03167 (2.36x) | 31.09 | 78.53 |
| +UnI-R (T=32) | 0.40749 | 0.01700 | 0.00187 | 1.28503 (1.49x) | 36.14 | 81.16 |
| +UnI-R (T=16) | 0.24886 | 0.01366 | 0.00185 | 1.12728 (1.31x) | 34.80 | 82.20 |
| +UnI-R (T=8) | 0.18029 | 0.01003 | 0.00189 | 1.04232 (1.21x) | 33.62 | 80.63 |
Future frame tokens are generated in parallel, and UnI-R directly uses their hidden states without decoding them into images. The additional operations are lightweight, with ACDC and F2R together taking only 18.87 ms at T=32. Compared with competing methods, it achieves higher accuracy and lower latency.
UnI-R enables flexible trade-offs between accuracy and cost, as shown in Figure 4 and Table 10 of orginal paper. With generation budget T=16, UnI-R achieves a V* score of 82.20 (+6.28) at1.31x latency, while T=8 retains a 4.71 improvement at 1.21x latency. These results demonstrate UnI-R's practical efficiency and flexibility and can be selected according to accuracy, latency, and memory requirements. We anticipate that continued advances in generation acceleration will further improve UnI-R's efficiency and scalability within unified multimodal models.
Q2.
How does ACDC compare with random selection at the same budget, and how does the same bridge perform when reading current-image features instead of future states?
We thank the reviewer for the comment. To evaluate the contribution of ACDC, we compare it against random selection under the same token budget. ACDC is designed to identify future states that provide complementary information beyond the observed scene. As shown below, replacing ACDC with random selection consistently degrades performance across all five benchmarks and still outperforms the baseline, validating the effectiveness of its coverage-driven selection strategy and UnI-R framework.
Rebuttal Table 4. Comparison of ACDC with random selection and future-state features with current-image features.
| Method | BabyVision | VTBench | V* | MME-RealWorld | SpatialEval |
|---|---|---|---|---|---|
| Cosmos3 | 8.51 | 23.87 | 75.92 | 44.40 | 64.69 |
| Full UnI-R | 17.27 | 31.29 | 81.16 | 52.74 | 68.54 |
| ACDC → Random Selection | 14.69 | 27.42 | 78.01 | 49.68 | 67.14 |
| Reading Current-Image Features | 9.02 | 21.94 | 76.44 | 45.03 | 64.90 |
We also evaluate the same F2R bridge using current-image features instead of imagined future states. The results show that simply reading the current visual features repeatedly provides little benefit over the backbone and performs substantially worse than using imagined futures. This suggests that repeatedly accessing existing visual information does not provide the complementary evidence needed for reasoning and may introduce redundant or distracting representations. In contrast, imagined future states provide additional information about potential scene evolution, which can be effectively integrated through F2R.
These experiments validate both the motivation behind ACDC and the importance of providing F2R with imagined future states instead of repeatedly reading static visual features.
A Gentle Follow-up on Our Responses
Dear Reviewer XNjd,
We hope this message finds you well. We sincerely appreciate your time and the thoughtful feedback you have provided. As the discussion period is nearing its end, with less than two days remaining, we would like to kindly ask whether our responses have adequately addressed your concerns. If any points remain unclear or warrant further discussion, we would be grateful for the opportunity to clarify them.
Thank you again for your time, consideration, and valuable feedback.
Best regards,
The Authors of Paper ''Understand, Imagine, then Reason: Thinking with Foresight in Unified Latent Space''
Response to Reviewer XNjd (Part 1)
We greatly appreciate Reviewer’s insightful comments. We address the concerns in detail below. If you have any further questions, please feel free to raise them, and we will be glad to address them.
W1.
We thank the reviewer for raising this question. We would like to clarify that Figure 4 and Table 10 of the original paper already examine the contribution of future content by varying the number of future frames at inference using the same fully trained UnI-R model. Disabling future generation (T=0) leaves only limited gains over the backbone. Figure 5 and Appendix Figure 7 additionally provide 12 qualitative examples illustrating how imagined futures inform the model’s answers.
To further address this concern, we have added controlled removal and replacement experiments. For removal, we first generate a fixed T=32 rollout and progressively remove its future representations. As shown in Table below, removing all future content brings performance close to the backbone across all five benchmarks, while retaining imagined futures restores the gains.
Rebuttal Table 1. Impact of future-content removal.
The replacement tests further examine whether the content of these representations matters. With the same UnI-R model and T=32 setting, we apply Gaussian perturbations, disrupt the ordering of future states, or replace them with futures from other samples, random vectors, or current-image features. Every intervention reduces performance on all five benchmarks relative to intact futures. The results support the importance of imagined futures specific to the input.
Rebuttal Table 2. Impact of future-content replacement.
The removal tests establish the contribution of future representations, while the replacement tests show that the gains depend on their content. The quantitative controls and qualitative examples demonstrate that UnI-R successfully feeds imagined future information back into the reasoning process at inference.